Home

Matrix Diffusion Toolkit User`s Manual - CLU-IN

1. Initial 20 ft yr Calibrated 25 ft yr PCE Initial 60 mg L Calibrated 143 mg L Initial 8 2E 10 m sec Calibrated 4 2E 10 m2 sec 0 33 1 33 1 7 g mL 1 7 g mL 0 0596 Initial 0 196 Calibrated 0 1596 155 L kg 0 0003 ft 56 ft 102 ft 1962 2011 Y USER S MANUAL Y Site information Literature Toolkit default Site information Literature Toolkit default Estimated site information Site information Initial maximum observed Calibrated PCE solubility Literature Toolkit default Calibrated value within the range of values reported in literature Literature Toolkit default Literature Toolkit default Literature Toolkit default Literature Toolkit default Estimated site information Estimated site information Calibrated within range of estimated site values Literature Toolkit default Literature Toolkit default Estimated as 50 of building length Estimated as 50 of building width Estimated site information Assumed continuous source Field data Site map distance of OU3 3 from the source Field data 146 CASE STUDY 3 FORMER DRY CLEANER FLORIDA DANDY SALE MODEL DSM Summary e The Toolkit DSM was used to estimate PCE soil concentrations in the low k zone at a former dry cleaner site OU 3 Building 106 at Naval Air Station Jacksonville Florida studied using Uni
2. HYDROGEOLOGY Transmissive Zone Description English Units Transmissive Zone Effective Porosity Sand X 0 35 Low k Zone Description Low k Zone Total Porosity Transmissive Zone Seepage Velocity TRANSPORT Key Constituent enter directly or choose from drop down list Plume Loading Concentration Immediately Above Low k Zone in Vertical Plane Source During Loading Period Molecular Diffusion Coefficient in Free Water Transmissive Zone Apparent Tortuosity Factor Exponent Low k Zone Apparent Tortuosity Factor Exponent Bulk Density of Transmissive Zone Bulk Density of Low k Zone Distribution Coefficient or Transmissive Zone Fraction of Organic Carbon Low k Zone Fraction of Organic Carbon Organic Carbon Partitioning Coefficient Next Step Show Graph Silt 0 43 3 70E 01 ma v Calculate V 2 TCE TCE z 1100 m 9 10E 10 od E 0 33 0 33 1 70 g mL 1 50 5 40E 04 9 33E 01 3 80E 04 Show Previous Results g mL mL g Calculated R o 1 17 8 1 18 L kg New Site Clear Data Save Data Load Data Return to Model Selection GENERAL CONNECTICUT DATA INPUT INSTRUCTIONS TTT Enter value directly Value calculated by Toolkit Do not enter data DNAPL Transmissive Zone Source Low k Zone Results Calculated Here SOURCE ZONE CHARACTERISTICS Source Zone Leng
3. To see matrix diffusion impacts in a downgradient plume The Black Box is drawn around the highest concentration contour downgradient of the source area The Blue Box is drawn around the second highest concentration contour downgradient of the source area To see matrix diffusion impacts downgradient of a Permeable Reactive Barrier PRB The Black Box is drawn around the highest concentration contour downgradient of the PRB The Blue Box is drawn around the second highest concentration contour downgradient of the PRB The width of the box is the width of the PRB Both models assume a two layer configuration where a plume in a transmissive zone is in contact with a low k zone The loading period where contaminants diffuse from the transmissive zone into the low k zone has to be estimated followed by a release period where contaminants diffuse from the low k zone into the transmissive zone One of the key challenges for running the Toolkit is coming up with good estimates for the year the loading period started and year the release period started In addition a loading concentration is required to run the model This is the concentration in the modeled area the boxes described on the previous page from the time the source started until the loading period ended This is often before the time any groundwater monitoring wells were installed and determining this value can be diff
4. DSM RESULTS DSM Model Results NOTE Due to run time constraints the DSM does not automatically produce results over multiple times The user can manually run the model for various times as necessary PARAMETER SEE 2 D LOW K AQUEOUS CONC Description Output showing the low k zone aqueous concentrations along the lateral distance from the source as a function of depth in the low k zone PARAMETER SEE LOW K AQ CONC VS DIST Description Concentration vs distance in the low k zone The user may use the Log 2Linear button to see the results on a semi log plot PARAMETER SEE LOW K AQ CONC VS DEPTH Description Concentration vs depth in the low k zone The user can vary the lateral distance from source at which to view results by 1 Entering the distance and 2 Then pressing the Update Graph button The user may use the Log 2Linear button to see the results on a semi log plot PARAMETER e SEE 2 D LOW K SORBED CONC Description Output showing the low k zone sorbed concentrations along the lateral distance from the source as a function of depth in the low k zone PARAMETER si SEE 2 D LOW K TOTAL CONC EE Output showing the low k zone total concentrations along the lateral distance from the source as a function of depth in the low k zone PARAMETER SEE TRANS ZONE AQUEOUS CONC Description Aqueous phase concentration vs distance in the transmissive zone Concentration is calculated by assuming a 10 foot
5. Standard Method Enter your Representative Concentration loading concentration C51 directly for this first area using Data Source 1 or Data Source 2 Contour Map Method The Toolkit calculates the geometric mean of the highest historical concentration within the black box and the contour line representing the black box This is the Representative Concentration during the loading period abbreviated C44 You can override this value if you want and just enter what you think is a good historical loading concentration for the black box area Step 5 7 Standard Method Skip this Step all you need is the Representative Concentration Cs2 Step 5 6 Contour Map Method Determine the concentration of the next highest concentration contour line from Step 5 3 above If the highest contour line is 100 mg L then use the 10 mg L contour Step 5 8 Standard Method Enter your Representative Concentration loading concentration C52 directly for this second area using Data Source 1 or Data Source 2 If you don t want to use this second area just set the concentration equal to the black box in the SRM input screen Contour Map Method The Toolkit calculates the geometric mean of the highest historical concentration within the blue box and the contour line representing the blue box this is the Representative Concentration during the loading period abbreviated C 2 You can override this value if you want and just
6. The Toolkit is primarily designed for unconsolidated sites with two layers a transmissive zone and a low k zone Although it can be used for fractured rock sites the application and interpretation will require additional interpretation and expertise The model basically assumes a single transmissive zone which would be a fracture and a single low k zone the rock matrix To apply this to a fractured system the mass discharge and concentration would have to be multiplied by two to account for the contribution from both sides of the fracture To simulate multiple fractures you would have to multiply the results from a single fracture by the number of fractures contributing to the mass flux mass discharge at the point of interest What contaminants can be modeled with the Toolkit To date most of the research involving matrix diffusion for low k zones has focused on chlorinated solvents such as TCE trichloroethene and methyl tert butyl ether MTBE However in theory matrix diffusion processes should apply to almost any dissolved contaminant including benzene and the other aromatic compounds found in gasoline although the overall impacts may differ Matrix diffusion of dissolved metals and radionuclides could also be modeled if a simplifying assumption of linear sorption desorption relationship and no degradation can be applied Can the Toolkit be used at LNAPL sites In theory many of the processes at chlorinated solvent sites will be
7. Typical Values 0 0002 0 02 for transmissive zones Source of Data The fraction organic carbon value should be measured if possible by collecting a sample of aquifer material from an uncontaminated saturated zone and performing a laboratory analysis for transmissive zones eg ASTM Method 2974 87 or equivalent If unknown a default value of 0 001 is often used e g ASTM 1995 How to Enter Data Enter directly PARAMETER LOW k ZONE FRACTION ORGANIC CARBON f Description Fraction of the aquifer material comprised of natural organic carbon in uncontaminated areas More natural organic carbon means higher adsorption of organic constituents on the aquifer matrix Typical Values Although based on limited data 0 0002 0 10 for low K zones is a likely range But some sites may be higher or lower Examples At the Moffatt Field site the foc of the clay fraction is about 0 0066 Roberts et al 1990 Domenico and Schwartz 1990 report these values silt Wildwood Ontario 0 00102 from Oconee River sediment coarse silt 0 029 medium silt 0 02 fine silt 0 0226 Chapman and Parker 2005 report a foc of glaciolacustrine aquitard composed of varved silts and clays 0 0024 to 0 00104 with an average of 0 00054 Adamson 2012 reports foc 0 001 for a clay layer in Jacksonville Florida and foc values for silts at the MMR site in Massachusetts ranging from lt 0 0005 to 0 0022 median value 0 0014 for one core usi
8. SQUARE ROOT MODEL SRM Summary e The Toolkit SRM was used to estimate bromide and fluorescein tracer groundwater effluent concentrations from a sand tank For fluorescein input parameters are shown on Figures 2 2 and 2 3 and comparisons of simulated and observed concentrations on Figure 2 4 For bromide input parameters are shown on Figure 2 5 and output on Figure 2 6 e SRM Plume Analysis model Section 2 was used to estimate the groundwater concentrations e Site hydrogeological data was entered in Section 3 transport parameters in Section 4 plume characteristics in Section 5 and source loading information in Section 6 e Site specific values as documented by Chapman et al 2012 were used for all parameters s An uncertainty of a factor of 10 was assumed for concentration estimations e To account for the flushing time in the tank model output from Day 25 the end of the loading period were compared to tank data from Day 32 the end of the loading plus 7 days of flushing of the transmissive zone This allowed for a pure diffusion vs diffusion comparison between model and tank data e Monitoring data from the sand tank study was used for calibration e The SRM assumes diffusion into and from the top interface of a single low k layer To account for the four distinct clay layers and associated eight interfaces in the sand tank the SRM output concentrations were multiplied by eight e Note that the Toolkit
9. Sand tank Low k Zone Mean concentration 1 mg L Sand tank Molecular diffusion coefficient 5 5E 10 m sec Literature sand tank study in free water Trans zone apparent 1 Literature sand tank study tortuosity factor exponent Low k zone apparent tortuosity factor exponent 1 Literature sand tank study e Trans zone bulk density 1 7 g mL Literature Toolkit default Low k zone bulk density 1 7 g mL Literature Toolkit default Retardation factor 1 Literature sand tank study Organic carbon partitioning 93 3 L kg Literature coefficient Transverse Vertical 0 001 m Literature Toolkit default hydrodynamic dispersivity Source Zone Source zone length 10000 m Assumed to account for the Characteristics extremely thin clay layers Source zone width 0 03 m Based on area of lengths of clay layers in sand tank and Source loading starts in year 2006 width of tank Source removed in year 24 days Sand tank study Sand tank study General See results for year 30 62 89 124 days Lateral distance from source 0 71 m Vertical depth 0 001 m Sand tank study MATRIX DIFFUSION TOOLKIT v USER S MANUAL Y 142 CASE STUDY 2B SAND TANK STUDY DANDY SALE MODEL DSM Summary The Toolkit DSM was used to estimate fluorescein groundwater concentrations in the low k zone in a sand tank Toolkit inputs are shown on Figure 2 7 Th
10. This is the typical historical concentration in the modeled area the boxes described on the previous page from the time the source started until the loading period ended This is often before the time any groundwater monitoring wells were installed We provide two data sources and two methods that can be used to obtain loading concentrations Data Source 1 Site History or Process Information Some sites might have available certain process knowledge about the modeling area during the loading period such as this area had DNAPL or there was a release of a certain strength waste In this case estimate the historical groundwater concentrations based on this information such as the effective solubility of the contaminant in a DNAPL and use this as the Loading Concentration For example the effective solubility of a constituent in a known DNAPL pool in the source could be used when modeling the source zone or if the DNAPL in the pool was comprised of 5096 Trichloroethene TCE a concentration of 550 mg L 50 of TCE solubility of 1100 mg L could be used Alternatively one could use an estimate of the average historical concentration from the time the source started to the end of the loading period sometimes a groundwater model with a source decay term such as REMChlor Falta et al 2007 can be used to estimate historical groundwater concentrations in the early period of a plume s life cycle Data Source 2 Highest Observed Concentration
11. 14 Compartment Model a discussion on the quantitative application of the 14 Compartment Model Sale et al 2008a The Matrix Diffusion Toolkit was developed for the ESTCP by GSI Environmental Inc Houston Texas in conjunction with Colorado State University Fort Collins Colorado MATRIX DIFFUSION TOOLKIT v USER S MANUAL Y 3 INTENDED USES FOR MATRIX DIFFUSION TOOLKIT AND LIMITATIONS The Matrix Diffusion Toolkit attempts to assist site managers and site consultants better understand matrix diffusion and help site stakeholders determine if matrix diffusion processes are significant enough to cause rebounding of downgradient plume concentrations above remediation goals after plume remediation or isolation is complete Having this information readily available before a remedy is implemented could assist site stakeholders select more appropriate remedies and improve effective risk communication with regulators and the public The Toolkit is intended to be used in two ways 1 As a screening level tool for simulating matrix diffusion effects The Toolkit brings key technical resources easy to use calculation worksheets and case studies together into one easy to access platform a In addition the Toolkit provides two methods for analyzing uncertainty in the estimation of mass discharge concentration and mass using the Square Root Model module One option performed automatically provides a lower range mostly likely value and
12. Falta et al 2007 can be used to estimate historical groundwater concentrations in the early period of a plume s life cycle Data Source 2 Highest Observed Concentration More commonly good Data Source 1 information will not be available In that case we recommend using the highest observed concentration from a groundwater monitoring point in the modeled area the two boxes and a groundwater concentration contour map from the period with the highest observed concentrations from the monitoring network This is typically the oldest concentration contour map available While not perfect this method is based on real data and represents observed loading concentrations in the modeled area Contour Map Method This is calculated automatically by the model from the previous data above as the geometric mean of highest historical concentration and the contour line concentration You can override this value How to Enter Data Enter directly or let the Toolkit calculate it Note that if overwritten the Toolkit calculated value in the blue cell can be replaced by pressing the Restore button PARAMETER NEXT HIGHEST CONCENTRATION ZONE APPROXIMATE LENGTH L2 Description Standard Method You can model two separate areas and the Toolkit will combine the diffusion processes Enter the length of your second modeled area or leave blank if you only want to model one area Contour Map Method Concentration of contour line denoted by the blue box i
13. H Wilson J T Kampbell D H Miller R N and Hansen J E 1995 Technical Protocol for Implementing Intrinsic Remediation with Long Term Monitoring for Natural Attenuation of Fuel Contamination Dissolved in Groundwater Revision 0 Air Force Center for Environmental Excellence Wiedemeier T H H S Rifai C J Newell and J T Wilson 1999 Natural Attenuation of Fuels and Chlorinated Solvents in the Subsurface Wiley New York pp 615 MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 75 MATRIX DIFFUSION TOOLKIT TROUBLESHOOTING TIPS Minimum System Requirements The Matrix Diffusion Toolkit model requires a computer system capable of running Microsoft Excel 2007 or 2010 for Windows Operation requires an IBM compatible PC equipped with a Pentium or later processor running at a minimum of 450 MHz A minimum of 256 MB of system memory RAM is strongly recommended Computers not meeting these recommendations will experience slow running times and or problems with memory Installation and Start Up The software is installed by unzipping the Toolkit model file MatrixDiffusionToolkit zip and keeping all the unzipped files in the same folder on your computer hard drive To use the software start Excel and load the MatrixDiffusionToolkit xlsm model file from the File Open menu If you are using Excel 2010 you may see a message box that asks you whether you want to disable or enable the macros For the Toolkit to operate e
14. Number of Interfaces 1 Plume in transmissive zone in contact with 2 interfaces but top one is very thin and can t store much mass Don t count top low k layer Number of Interfaces 1 Plume in transmissive zone in contact with 2 interfaces both low k units gt 1 meter thick Count both interfaces Number of Interfaces 2 Plume of same concentration in transmissive zone in contact with 3 interfaces all low k units 1 meter thick Number of Interfaces 3 Use Toolkit results with no adjustment Use Toolkit results with no adjustment Multiply all Toolkit results by 2 Multiply all Toolkit results by 3 MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 10 FREQUENTLY ASKED QUESTIONS How accurate are the results The two models utilized in the Toolkit are very simplified representations of an extremely complicated process and field conditions Therefore even with sampling data from the low k zones we consider the potential results as an order of magnitude range accuracy But at many sites this level of accuracy will still provide very useful information for site managers What input data will need Some of the input data are similar to what is used for existing solute transport models e g Darcy groundwater velocity size of the modeled area information on when the source started etc Other input data will look new to many users for example you ll need to es
15. a release of contaminants from the low k zone Assumptions The model makes the following assumptions 1 A source considered to be a thin pool is introduced at the contact between the two layers upgradient of x 0 2 A loading period occurs where there is a constant concentration of contaminants in the transmissive zone that drives contaminants into the low k zone 3 A release period occurs where the transmissive zone is assumed to have no concentration and an upper range estimate of release from the low k zone is generated 4 There is no degradation in either layer 5 Both layers are uniform homogeneous isotropic and infinite in the z direction perpendicular to groundwater flow 6 One dimensional 1 D advective transport in the transmissive layer parallel to the boundary of the layers is accompanied by transverse dispersion and diffusion 7 There is no longitudinal dispersion in the transmissive layer 8 1 D transverse diffusion transport occurs in the low k layer 9 Retardation of contaminants in both layers is based on instantaneous equilibrium between aqueous and sorbed phases Summary Sorbed phase mass in the transmissive layer at any time t can be calculated as M t Mag t n where M t Sorbed phase mass in the transmissive layer at any time t M M t Aqueous phase mass in the low k layer at time t M calculated using Appendix A 2 10 n Porosity of transmissive layer unit
16. diffusion modeling in general we consider the two models in the Toolkit to be an order of magnitude a factor of 10 level of accuracy tools While this seems a large range the results will provide useful information in context of the wide range of concentrations and mass discharge found in source zones for example see the paper Contaminant Plume Classification System Based on Mass Discharge by Newell et al 2011 This level of accuracy means that there is probably no need to spend considerable effort in trying to calibrate the models to the 2 or 3 significant digit MATRIX DIFFUSION TOOLKIT v USER S MANUAL Y 19 SRM DATA ENTRY SRM Data Input Screen Step 1 System Units PARAMETER SYSTEM UNITS Unit system to perform matrix diffusion calculations in SI System meters etc or English Units feet etc How to Enter Data Choose the appropriate radio button Step 2 Analysis Type PARAMETER ANALYSIS TYPE Description Type of matrix diffusion analysis to perform Select Source Zone Analysis to see matrix diffusion impacts in a source zone Select Plume Analysis to see matrix diffusion impacts in a downgradient plume Select PRB Analysis to see matrix diffusion impacts downgradient of a PRB How to Enter Data Choose the appropriate radio button MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 20 SRM DATA ENTRY Step 3 Hydrogeology PARAMETER LOW k ZONE DESCRIPTION Description of the l
17. typically silt or clay into a high permeability advection dominated unit typically sand or gravel Estimates of concentration and mass remaining in the high permeability unit after the source is removed are also provided 2 Dandy Sale Model A module allowing users to perform 1 contaminant transport via advection and transverse diffusion in the transmissive layer and 2 transport via transverse diffusion in the low k zone The module provides planning level estimates of Low k Zone i Aqueous sorbed and total concentration and MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 2 INTRODUCTION ii Aqueous sorbed and total mass Transmissive Zone i Aqueous sorbed and total concentration ii Aqueous sorbed and total mass and iii Mass discharge 3 Matrix Diffusion Related Tools An additional feature that provides a review of theory and methods related to matrix diffusion a NAPL Dissolution Calculator a module that estimates the transverse diffusion of contaminants into the groundwater passing over the top of a Non aqueous Phase Liquid NAPL pool and the transverse diffusion of contaminants into the low k unit underlying the pool b Plume Magnitude Information a summary of the Plume Magnitude Classification System Newell et al 2011 and its application to site investigation and remediation c Low k Zone Remediation Alternatives a summary of current alternatives for the remediation of low k zones and d
18. 06 cm s TRRP 2008 Note that there is a wide range of reported values for example Wiedemeier et al 1999 report a D for benzene of 1 1E 05 cm s For more information see Pankow and Cherry 1996 for solvents and Wiedemeier et al 1999 variety of constituents Source of Data Chemical reference literature such as Pankow and Cherry 1996 for solvents Wiedemeier et al 1999 variety of constituents or other references with chemical properties How to Enter Data 1 Select units and 2 Enter directly Note that if the constituent is selected from the drop down list the Toolkit provides a default value for the parameter PARAMETER TRANSMISSIVE ZONE APPARENT TORTUOSITY FACTOR EXPONENT p Description The Apparent Tortuosity Factor x relates the molecular diffusion coefficient in free water Do of a constituent in a porous medium to its effective diffusion coefficient De Values of t can range between 0 and 1 Estimations of t can be obtained using the relationship TE P Where 6 is the porosity and p the Apparent Tortuosity Factor Exponent Depending on the geologic medium values for p can vary between 0 3 and 5 4 Charbeneau 2000 Pankow and Cherry 1996 Dullien 1992 Lerman 1979 and Millington and Quirk 1961 Typical Values Sand 0 33 Silt 0 33 Payne et al 2008 How to Enter Data Enter directly Note that if the transmissive zone description is selected from the drop down list the Tool
19. 114 CASE STUDY 2 SAND TANK STUDY asses eee sese ereer eenn 133 CASE STUDY 3 FORMER DRY CLEANER FLORIDA eese 145 MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y iii QUICK START Matrix Diffusion Low k zones i e low permeability zones such as silt clay layers can serve as indirect low level sources of contamination to transmissive zones due to matrix diffusion If you can apply several simplifying assumptions about heterogeneity at your site coupled with its concentration history the Matrix Diffusion Toolkit Toolkit can provide planning level estimates of e mass discharge sometimes called mass flux in grams per day and or concentrations in the transmissive zone caused by matrix diffusion and e mass of contaminants and concentrations in the low k zone Why is this important Understanding and evaluating matrix diffusion can provide information regarding a variety of key questions such as 1 If remediate a transmissive zone but my remediation technology doesn t remove contaminants from low k zones in contact with the transmissive zone will be able to achieve my cleanup standards How much mass could be present in low k zones at my site If install a permeable reactive barrier will have trouble achieving cleanup standards downgradient of the barrier 4 Ifl remove all the DNAPL is there a chance lll still be above MCLs How much longer might have to wait for a source zon
20. 2000 Groundwater Hydraulics and Pollutant Transport Prentice Hall Upper Saddle River New Jersey Cohen R M and J W Mercer 1993 DNAPL Site Evaluation CRC Press Boca Raton FL Davis S N 1969 Porosity and Permeability of Natural Materials in De Wiest R J M ed Flow Through Porous Media New York Academic Press p 53 89 Domenico P A and F W Schwartz 1990 Physical and Chemical Hydrogeology Wiley New York New York Dullien F A L 1992 Porous Media Fluid Transport and Pore Structure 2nd edition 574 pp Academic San Diego California 1992 Falta R W M B Stacy A N M Ahsanuzzaman M Wang and R C Earle 2007 REMChlor Remediation Evaluation Model for Chlorinated Solvents User s Manual U S Environmental Protection Agency Center for Subsurface Modeling Support Ada OK September 2007 Farhat S K P C de Blanc C J Newell J R Gonzales and J Perez 2004 SourceDK Remediation Timeframe Decision Support System User s Manual Developed for the Air Force Center for Engineering and the Environment AFCEE by GSI Environmental Inc Houston Texas http www gsi net com en software free software sourcedk html Gelhar L W C Welty and K R Rehfeldt 1992 A Critical Review of Data on Field Scale Dispersion in Aquifers Water Resources Research 28 7 1955 1974 Johnson A and DA Morris 1962 Physical and hydrologic properties of water bearing deposits from core holes in the Los Ba
21. Figure 1 Advancing solvent plume g issi 3 Low permeability silts Transmissive sand Simultaneous inward and outward diffusion in stagnant zones Figure 1 Conceptual model of matrix diffusion effects as part of plume response Source T Sale T Illlangasekare AFCEE 2007 The potential for matrix diffusion effects can be seen at virtually any site with heterogeneity in the subsurface dense non aqueous phase liquid DNAPL and or where persistent groundwater contaminant concentrations after source zone remediation have been observed While matrix diffusion has been identified as a potential problem there are relatively few tools available to help practitioners in the field determine if matrix diffusion could be a problem at their site Currently the field methods are still based on research techniques that are relatively expensive e drilling collecting soil samples etc There are site factors e high heterogeneity low groundwater flow rate high contaminant solubility etc which can be evaluated to qualitatively estimate if matrix diffusion effects are expected to be significant However current analytical fate and transport models such as BIOCHLOR and REMChlor or complex numerical models such as MODFLOW MT3D cannot accurately simulate matrix diffusion effects Some simple equations have been developed as part of an Air Force Center for Engineering and the Environment AFCEE research project AFCEE 2007 that
22. In cases where a good comparison between concentrations and or mass discharge from actual groundwater monitoring data can be made either because there is no residual source or the matrix diffusion signal can be abstracted out the recommended sequence of model input values to change is MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 18 MATRIX DIFFUSION TOOLKIT MODELS a First change the representative concentrations Cs and Cs2 in the black box and blue box respectively If the simulated concentrations are higher than observed concentrations reduce the representative concentrations in the black and blue boxes b If it is still difficult to get a good fit try changing either the start or end of the loading period if there is some uncertainty on the exact years of these two times To increase the simulated concentration move the start of the loading period back in time or the start of the release period more recent in time In other words more time for diffusion during the loading period will result in higher concentrations during the release periods c To further improve the match after working with the previous two steps consider changing some of the hydrogeologic and or transport properties such as Darcy velocity low k zone tortuosity and low k zone retardation factor Other parameters in the model can also be changed to develop a better match Because of the simplifying assumptions in the model and the early state of matrix
23. MODEL Mig tv Maq tv M tv D where ML t Total mass at time t after the source has depleted M ML d Aqueous phase mass at time t after the source has depleted M calculated using Appendix A 2 4 Equation 2 and M t 7 Sorbed phase mass at time t after the source has depleted M calculated using Appendix A 2 5 Equation 2 MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 97 APPENDIX A 2 DANDY SALE MODEL Appendix A 2 7 Transmissive Layer Aqueous Concentration Purpose Determine the transmissive layer aqueous concentration output in the Dandy Sale Model of the Matrix Diffusion Toolkit Note This derivation was originally developed by Sale et al 2008b and described in detail in Sale et a 2008b and Bolhari 2012 Given The vertical plane source produces a plume in the transmissive zone that loads up the low k zone due to diffusion This vertical plane source is shut off and diffusion results in a release of contaminants from the low k zone Assumptions The Toolkit uses a simplified conceptual model of a two layer aquifer system a transmissive layer above a low k layer Figure A 2 7 1 A source e g DNAPL is introduced at the contact between the two layers As shown in the figure x is in the direction of groundwater flow and z the direction perpendicular to groundwater flow The edge of the source at the interface between the two layers is designated x 0 and z 0 with both x and z increasing with di
24. Matrix Diffusion Toolkit was developed for the Environmental Security Technology Certification Program ESTCP of the U S Department of Defense by GSI Environmental Inc in collaboration with Colorado State University ESTCP Project Officer Matrix Diffusion Toolkit Code Developers Matrix Diffusion Toolkit User s Manual Dandy Sale Model Equations Developers Matrix Diffusion Toolkit Graphics Matrix Diffusion Toolkit Review Team Matrix Diffusion Toolkit Review Team Matrix Diffusion Toolkit Review Team Matrix Diffusion Toolkit Review Team Matrix Diffusion Toolkit Review Team Matrix Diffusion Toolkit Example Dataset Dr Andrea Leeson Dr Shahla Farhat Dr Charles Newell GSI Environmental Inc 2211 Norfolk Suite 1000 Houston Texas 77098 phone 713 522 6300 skfarhat gsi net com Dr Shahla Farhat Dr Charles Newell GSI Environmental Inc Houston Texas Dr David Dandy Dr Thomas Sale Ms Jennifer Wahlberg Colorado State University Fort Collins Colorado Christina Walsh GSI Environmental Inc Houston Texas V Yates J Small K Holzheimer J McDade Dr D Adamson Dr D Mackay GSI Environmental Inc Houston Texas J Wahlberg Colorado State University Fort Collins Colorado Dr R H Anderson AFCEE Lackland AFB Texas B J Holloway and C G Coyle USACE Omaha Nebraska Dr R Falta Clemson University Clemson South Carolina The example dataset used in
25. More commonly good Data Source 1 information will not be available In that case we recommend using the highest observed concentration from a groundwater monitoring point in the modeled area the two boxes and a groundwater concentration contour map from the period with the highest observed concentrations from the monitoring network This is typically the oldest concentration contour map available While not perfect this method is based on real data and represents observed loading concentrations in the modeled area Step 5 4 Standard Method Skip this Step all you need is the Representative Concentration Step 5 6 Contour Line Method Determine the loading concentration using Data Source 1 or Data Source 2 Use the maximum concentration from any well within the highest concentration contour denoted by the black box in the Toolkit input screen figure Your goal is to get a concentration that reflects historical conditions before the monitoring system was installed at most sites Step 5 5 Standard Method Skip this Step all you need is the Representative Concentration Cai Step 5 6 Contour Line Method Determine the concentration of the highest concentration contour denoted by the black box in the Toolkit input screen figure Step 5 2 above Again use a contour map from the highest concentration period where groundwater samples were collected MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 29 SRM DATA ENTRY Step 5 6
26. Organic Carbon Partitioning Coefficient Next Step Show Graph Show Previous Results Sand Sand X 0 25 Clay Clay 0 47 Calculate V 2 2 00E 01 PCE ms m2 sec el H o g mL g mL mL g Calculated R 5 00E 04 1 00E 03 1 55E 02 L kg New Site Clear Data Save Data Load Data Return to Model Selection FORMER DRY CLEANER GENERAL FLORIDA DATA INPUT INSTRUCTIONS TTT Enter value directly Value calculated by Toolkit Do not enter data DNAPL Transmissive Zone Source Low k Zone Results Calculated Here SOURCE ZONE CHARACTERISTICS Source Zone Length Source Zone Width Transverse Vertical Hydrodynamic Dispersivity Source Loading Starts in Year Source Removed in Year 102 ft 3 00E 04 ft 1962 format yyyy 2011 format yyyy Restore zl See Release Period Results for Year 2011 format yyyy Lateral Distance from Source 65 ft Depth into Low k Zone 16 5 ft we Return to Main Screen Figure 3 2 DSM Input Parameters Initial MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 148 CASE STUDY 3 FORMER DRY CLEANER FLORIDA DANDY SALE MODEL DSM Data Input Screen DATA INPUT INSTRUCTIONS TTT Enter value directly Matrix Diffusion Toolkit Version 1 0 Value calculated by Toolkit Do not enter data Site Location and ID Jacksonvil
27. all DNAPL is gone but SRM would simulate a more accurate source representation 5 want to know the concentration vs depth profile in a low k zone Dandy Sale Concentration vs depth plot or Concentration vs lateral distance plot 6 want to make sure the matrix diffusion model accounts for Dandy Sale Concentration vs time plot or contaminant concentrations in the transmissive zone when mass discharge vs time plot calculating the release from low k zones 7 want to account for the travel time of the plume in the transmissive Dandy Sale Concentration vs time plot or zone so that the loading period for the downgradient low k zones mass discharge vs time plot starts later than the loading period for the near source low k zones This is more important for plumes with long residence times 20 years Concentration assuming a monitoring well with a 10 foot screened interval this cannot be changed in the model MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 12 FREQUENTLY ASKED QUESTIONS What are the key input data for the Square Root Model in the Toolkit The Square Root Model originally based on work performed by Drs Beth Parker and John Cherry and modified by Dr Tom Sale asks you to provide these input data 1 What is the length and width of each zone You get to model two zones with different sizes and different concentrations See the Data Input Section for examples of how to determine
28. an upper range for estimated outputs based on the specified source area concentrations The second option Advanced Uncertainty Analysis utilizes a Monte Carlo type approach to analyze uncertainty in the actual concentration porosity apparent tortuosity factor exponent and retardation factor measurements With this tool groundwater practitioners can estimate the accuracy of the hydrologic measurements that are being used for the matrix diffusion calculation b The Toolkit can also be used to estimate the diffusion of contaminants into the groundwater passing over the top of a NAPL pool and the diffusion of contaminants into the low k unit underlying the pool 2 As a tool for learning about matrix diffusion The Toolkit reviews emerging methodologies associated with site characterization and matrix diffusion such as the 14 Compartment Model Sale et al 2008a and the Plume Magnitude Classification System Newell et al 2011 The Toolkit has the following assumptions and limitations e Assumes the user is familiar with basic groundwater transport and mass balance concepts e Uses a simplified conceptual model of a two layer aquifer system a transmissive layer and a low k layer where there are two different time periods o A loading period where there is a constant concentration of contaminants in the transmissive zone that drives contaminants into the low k zone and MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 4 INTENDED US
29. clay 0 37 Soft slightly organic clay 0 66 Soft very organic clay 0 75 Soft bentonite 0 84 One fractured microcrystalline limestone in Virginia had matrix porosities ranging from 0 0004 to 0 0065 GSI Environmental Source of Data Typically estimated Occasionally obtained through physical property testing of site soil samples How to Enter Data Enter directly Note that if the low k zone description is selected from the drop down list the Toolkit provides a default value for the parameter MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 21 SRM DATA ENTRY PARAMETER TRANSMISSIVE ZONE DARCY VELOCITY V4 cm sec ft or m day ft or m yr Description Transmissive zone groundwater Darcy velocity To characterize concentrations in a well with a 10 foot screened interval in the transmissive layer representative measurements are required for the Darcy velocity or both the hydraulic flow gradient and the hydraulic conductivity of the flow system Representative measurements of the Darcy velocity should be obtained at one or more locations using appropriate slug or pumping test methods In the SRM Darcy velocity is only used for calculation of concentration from the mass discharge output Typical Values 0 2 200 ft yr 0 06 61 m yr Newell et al 1996 Source of Data Calculated by multiplying hydraulic conductivity by hydraulic gradient Va K x i Use of actual site data for hydraulic conductivity and hydraulic gradient par
30. compared to observed tetrachloroethene PCE soil concentrations at three locations in the downgradient plume e Step 3 Input parameters were adjusted as needed to improve the comparison of simulated and observed PCE concentrations GROUNDWATER FLOW Figure 3 1 Site Layout Building 106 in Operable Unit 3 Naval Air Station Jacksonville Florida MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 145 CASE STUDY 3 FORMER DRY CLEANER DANDY SALE MODEL FLORIDA DSM Input Data Data Type Source of Data Hydrogeology Transport Low k Zone Source Zone Characteristics General Trans zone description Trans zone porosity Low k zone description Low k zone porosity Trans zone seepage velocity Key constituent Mean concentration Molecular diffusion coefficient in free water Trans zone apparent tortuosity factor exponent Low k zone apparent tortuosity factor exponent Trans zone bulk density Low k zone bulk density Trans zone fraction organic carbon Low k zone fraction organic carbon Organic carbon partitioning coefficient Coefficient of transverse hydrodynamic dispersion Source zone length Source zone width Source loading starts in year Source removed in year See results for year Lateral distance from source Vertical depth MATRIX DIFFUSION TOOLKIT sand 0 25 clay 0 47
31. concentration and mass remaining in the high permeability unit after source removal Governing equations and assumptions are provided in Appendix A 1 Guidelines for selecting key input parameters for the model are outlined in Square Root Model Data Entry For help on results see Square Root Model Results Dandy Sale Model DSM Through a 2003 AFCEE project Sale et a 2008b Dr David Dandy at Colorado State University developed an exact analytical solution for a two layer scenario shown in Figure 2 Key attributes of the model include contaminant transport via advection and transverse diffusion in the transmissive layer transport via transverse diffusion in the low k zone unique retardation factors for each layer unique contaminant degradation rates for each layer and an adjustable source term MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 15 MATRIX DIFFUSION TOOLKIT MODELS Plume of aqueous sorbed contaminants Semi infinite low permeability zone e g C silt Source Co at the contact decaying exponentially with increasing distance from the interface Figure 2 A conceptual model of the two layer scenario A Active source c at the contact decaying exponentially into the transmissive layer B Depleted source source strength 0 from Sale et al 2008b The theoretical basis for the module is
32. diffusion has occurred diffusion coefficients fraction organic carbon of the clays and silts being modeled etc The Toolkit provides default values and advice on how to pick the best value that represents your site conditions How is site data converted to a simple configuration that can be modeled with the Toolkit What concentrations do enter First you pick which of two separate diffusion models to run see Page 11 The Toolkit then guides you through how to set up the selected model For example to determine the modeled area length and width for the SRM you can either enter your own length and width directly or use the following method based on a historical contour map see SRM Data Entry Step 5 Contour Line Method Draw a downgradient transect line perpendicular to groundwater flow and an upgradient transect line perpendicular to groundwater flow to define the area you want to see results for from the Toolkit Here are three examples where you need to enter the length and width of the black box and blue box in the drawing MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y ll QUICK START Type of Problem to be Analyzed Using the Toolkit Black Box in Drawing Blue Box in Drawing To see matrix diffusion impacts in a source zone The Black Box is drawn around the highest concentration contour in the source area The Blue Box is drawn around the second highest concentration contour in the source area
33. enter what you think is a good historical loading concentration for the blue box area PARAMETER HIGH CONCENTRATION ZONE APPROXIMATE LENGTH L Description Standard Method You can model two separate areas and the Toolkit will combine the diffusion processes This is the length of your first modeled area Contour Map Method Length of the highest concentration contour line on a historical plume map between the upgradient and downgradient transects that represent your modeled area denoted by the black box in the Toolkit input screen figure Typical Values 0 3300 ft 0 1000 m Source of Data Standard Method Modeled area length for this first of two subareas Contour Map Method Contour map should be from the highest concentration period where groundwater samples were collected For example if concentrations have been decreasing use a concentration contour map from 1990 and not 2012 How to Enter Data Enter directly MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 30 SRM DATA ENTRY PARAMETER HIGH CONCENTRATION ZONE APPROXIMATE WIDTH W Description Standard Method You can model two separate areas and the Toolkit will combine the diffusion processes This is the width of your first modeled area Contour Map Method Width of the highest concentration contour line on a historical plume map between the upgradient and downgradient transects that represent your modeled area denoted by the black box in the Toolkit input sc
34. for the parameter MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 64 DISSOLUTION MODEL DATA ENTRY PARAMETER APPARENT TORTUOSITY FACTOR EXPONENT p Description The Apparent Tortuosity Factor t relates the molecular diffusion coefficient in free water Do of a constituent in a porous medium to its effective diffusion coefficient De Values of t can range between 0 and 1 Estimations of t can be obtained using relationship De T OD pr Where 6 is the porosity and p the Apparent Tortuosity Factor Exponent Depending on the geologic medium values for p can vary between 0 3 and 5 4 Charbeneau 2000 Pankow and Cherry 1997 Dullien 1992 Lerman 1979 and Millington and Quirk 1961 Typical Values Sand 0 33 Silt 0 33 Payne et al 2008 How to Enter Data Enter directly Note that if the transmissive zone description is selected from the drop down list the Toolkit provides a default value for the parameter MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 65 DISSOLUTION MODEL DATA ENTRY Step 4 Plume Characteristics How to Enter Data Enter directly Note that if the constituent is selected from the drop down list the Toolkit provides a default value for the parameter PARAMETER HEIGHT OF NAPL POOL H PARAMETER NAPL DENSITY Puapi MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 66 DISSOLUTION MODEL DATA ENTRY Typical Values 0 80 1 4 Source of Data From an analysis of representative NAPL
35. has a built in 1 yr transition period between diffusion into the matrix and release from the matrix Because the total experiment time of 120 days is less than 1 yr this transition time was temporarily changed to 1 day for this Case Study KEY POINTS The purpose of this evaluation was to determine if the SRM in the Toolkit could be applied to simulate a difficult problem four very thin layers in a system with advection As described in the Uses and Limitations Section page 4 the SRM model assumes a two layer system with one interface an infinitely thick low k zone and instantaneous flushing of the transmissive zone instantly changing from the loading period to the release period The tank study had four very thin low k zones ranging from 0 03 to 0 2 meters thick compared to a theoretical contaminant penetration depth into an infinite low k zone of 0 25 m Finally the tank had a relatively long flushing period 7 days compared to the total experiment time of 120 days Despite these differences from the assumed configuration of the SRM the end results show the model was able to match the actual data from the tank within an order of magnitude For fluorescein although the general shape of the observed concentrations was followed using tank specific values documented by Chapman et al 2012 simulated concentrations were consistently underestimated by about a factor of 3 A much better MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 13
36. linear sorption desorption relationship and no degradation can be applied Is the Toolkit able to simulate degradation in the low k zone Not at this time Numerical problems prevented a full implementation of the Dandy Sale Model with degradation Sale et al 2008b Consequently this version of the Toolkit assumes no degradation in the low k zone However we hope to incorporate this feature in future versions of the Matrix Diffusion Toolkit MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 11 FREQUENTLY ASKED QUESTIONS Which model in the Toolkit should choose the Square Root Model or the Dandy Sale Model Want the Following Information Which Model Output 1 Mass Discharge sometimes called mass flux data from a low k Square Root Mass discharge vs time plot zone to a transmissive zone in units of grams per day vs time both OR Dandy Sale past and future 2 How much mass could be present in low k zones at my site Square Root Mass in low k zone vs time plot OR Dandy Sale 3 If install a permeable reactive barrier will have trouble achieving Square Root Concentration vs time plot or downgradient cleanup standards OR Dandy Sale mass discharge vs time plot 4 If remove all the DNAPL in a source zone is there a chance I ll still Square Root Concentration vs time plot or be above MCLs How much longer might have to wait for a source OR Dandy Sale mass discharge vs time plot zone to achieve MCLs after
37. low level sources of contamination to transmissive zones due to matrix diffusion If you can apply several simplifying assumptions about heterogeneity and a site s concentration history to your site the Toolkit can tell you e the mass discharge sometimes call mass flux in grams per day leaving the modeled area due to release from the low k zones e the concentration of contaminants in a monitoring well with a 10 foot screen located in the transmissive zone in the downgradient portion of the modeled area e the average concentration of contaminant leaving the source zone assuming some minimum flow zone above the modeled low k zone e the mass of contaminants in the low k zone at any time and e the concentration of contaminants anywhere in the low k zone at any time What questions can I address with the Matrix Diffusion Toolkit The Toolkit can be used to provide information regarding a variety of questions such as 1 If remediate a transmissive zone but my remediation technology doesn t remove contaminants from low k zones in contact with the transmissive zone will be able to achieve my cleanup standards 2 How much mass could be present in low k zones at my site 3 If install a permeable reactive barrier will have trouble achieving downgradient cleanup standards 4 IfI remove all of the DNAPL is there a chance I ll still be above MCLs 5 How much longer might have to wait for a source zone to achieve MCLs
38. measuring dispersion in the field However simple estimation techniques based on the length of the plume or distance to the measurement point scale are available from a compilation of field test data 0 05 times the modeled length Aziz et al 2000 in this case the NAPL pool length Typically estimated using empirical relationships Enter directly MOLECULAR DIFFUSION COEFFICIENT IN FREE WATER D cm sec m sec A factor of proportionality representing the amount of substance diffusing across a unit area through a unit concentration gradient in unit time Benzene 9 8E 06 cm s Tetrachloroethene 8 2E 06 cm s Ethylbenzene 7 8E 06 cm s Trichloroethene 9 1E 06 cm s Toluene 8 6 06 cm s cis 1 2 Dichloroethene 1 1E 05 cm s Xylene 8 5E 06 cm s Vinyl Chloride 1 2E 05 cm s MTBE 9 4E 05 cm s 1 1 1 Trichloroethane 8 8E 06 cm s TRRP 2008 Note that there is a wide range of reported values for example Wiedemeier et al 1999 report a Do for benzene of 1 1E 05 cm s For more information see Pankow and Cherry 1996 for solvents and Wiedemeier et al 1999 variety of constituents Chemical reference literature such as Pankow and Cherry 1996 for solvents Wiedemeier et al 1999 variety of constituents or other references with chemical properties 1 Select units and 2 Enter directly Note that if the constituent is selected from the drop down list the Toolkit provides a default value
39. of the input cells ZVALUE is displayed in a number box The most common cause of this problem is that some input data are missing Double check to make certain that data required for your run have been entered in all of the input cells and all options have been selected Common Error Messages Unable to Load Help File The most common error message encountered with the Toolkit is the message Unable to Open Help File after clicking on a Help button Depending on the version of Windows you are using you may get an Excel Dialog Box a Windows Dialog Box or you may MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 76 TROUBLESHOOTING TIPS see Windows Help load and display the error This problem is related to the ease with which the Windows Help Engine can find the data file MatrixDiffusionToolkit chm Here are some suggestions in decreasing order of preference for helping WinHelp find it s f you are asked to find the requested file do so The file is called MatrixDiffusionToolkit chm and it was installed in the same directory folder as the Matrix Diffusion Toolkit model file MatrixDiffusionToolkit xlsm e Use the File Open menus from within Excel instead of double clicking on the filename or Program Manager icon to open the Matrix Diffusion Toolkit model file This sets the current directory to the directory containing the Excel file you just opened MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 77 ACKNOWLEDGEMENTS The
40. of this plane In MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 5 INTENDED USES AND LIMITATIONS other words the Dandy Sale model only models matrix diffusion downgradient of a source zone e Concentration results from both the Square Root Model and Dandy Sale Model are based on estimates of mass discharge leaving the low k zone Concentrations are then calculated by assuming a 10 foot screened interval The 10 foot screened interval was selected because at an actual field site contamination diffusing from a low k zone might spread vertically above a 1 foot screen It was thought to be very unlikely that there would be more than 10 feet of vertical spreading in the transmissive zone Bottom line the 10 foot screened interval is hard wired into the models and cannot be changed by the user MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 6 FREQUENTLY ASKED QUESTIONS Why is matrix diffusion important Won t the DNAPL take so long to go away that matrix diffusion will never be that important Matrix diffusion can be a key process both at sites where remediation has not been conducted and at sites where much of the DNAPL has been removed by active remediation projects e For the no remediation case a simple modeling study of a 675 kg DNAPL release showed that it would take about 39 years for the DNAPL to dissolve away naturally and then it would take another 87 years until matrix diffusion went below a certain source strength 0 1 grams p
41. the buttons will function Additionally REMOVING OR ADDING rows or columns in input screens may cause the program to crash 3 Parameters used in the model are to be entered directly into the white blue cells NOTE Although literature values are provided site specific hydrogeologic transport and plume characteristic values will likely provide better results If literature values are used and there is uncertainty in the value chosen sensitivity analyses should be conducted to determine the effects of the uncertainty on model predictions MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 61 DISSOLUTION MODEL DATA ENTRY NAPL Dissolution Model Data Input Screen This module calculates the dissolution rate from the top of a DNAPL pool bottom of a LNAPL pool but not through the NAPL pool therefore the dissolution rate is likely underestimated Step 1 System Units PARAMETER SYSTEM UNITS Unit system to perform calculations in SI System meters etc or English Units feet etc How to Enter Data Choose the appropriate radio button Step 2 Hydrogeology Transmissive Zone PARAMETER UNIT DESCRIPTION Description of the transmissive zone How to Enter Data Choose from drop down list or enter directly PARAMETER POROSITY Description Dimensionless ratio of the volume of voids to the bulk volume of the surface soil column matrix Note that total porosity is the ratio of all voids including non connected voids to the bulk
42. the length and width of the two zones 2 How long years was the loading period when concentrations in the transmissive zone were higher than the low k zones This is based on your understanding of site history such as the time from the initial release to the time when remediation was or will be performed on the transmissive zone 3 What was the concentration during the loading period It is rare to have monitoring data from the time of the release to now so we ve provided some guidance based on the maximum concentration ever observed in the zones you are modeling see SRM Data Entry Step 5 4 How long years has release from low k zones been occurring In other words how long has it been since the transmissive zone concentration was lower than the concentrations that have diffused into the low k zones For sites where remediation has or will occur this is easy just enter the date when remediation reduced the concentrations in your modeling zones For other situations you can make some estimates to get an idea of the impact of matrix diffusion 5 What is the diffusion coefficient for the contaminant of interest The Toolkit provides a library of diffusion coefficients for the most common contaminants we deal with at sites 6 What are the key transport properties of the clay tortuosity and retardation factor The Toolkit provides a calculator for you to estimate these parameters if you are not familiar with them What are th
43. the start or end of the loading period if there is some uncertainty on the exact years of these two times To increase the simulated concentration move the start of the loading period back in time or the start of the release period more recent in time In other words more time for diffusion during the loading period will result in higher concentrations during the release periods 3 To further improve the match after working with the previous two steps consider changing some of the hydrogeologic and or transport properties such as Darcy velocity low k zone tortuosity and low k zone retardation factor Other parameters in the model can also be changed to develop a better match Because of the simplifying assumptions in the model and the early state of matrix diffusion modeling in general we consider the two models in the Toolkit to be an order of magnitude a factor of 10 level of accuracy tools Therefore there is probably no need to spend considerable effort in trying to calibrate the models to the 2 or 3 significant digit While this seems a large range the results will provide useful information in context of the wide range of concentrations and mass discharge found in source zones for example see the paper Contaminant Plume Classification System Based on Mass Discharge by Newell et al 2011 MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 45 DSM DATA ENTRY DSM Data Input Screen Step 1 System Units PARAMETER SYSTEM UNIT
44. time usually it is better to enter the earliest year How to Enter Data Enter directly PARAMETER SOURCE REMOVED IN YEAR Description Year source was removed This is either 1 the year that best represents when concentrations in the middle of the modeled area were reduced significantly by source remediation or 2 when source zone natural attenuation processes reduced the concentrations in the middle of the modeled area significantly For example the source could likely be considered removed by natural attenuation for the purposes of this model if the transmissive zone of the modeled area have been reduced by 9096 or 99 compared to the historical all time concentrations How to Enter Data Enter directly PARAMETER TRANSVERSE VERTICAL HYDRODYNAMIC DISPERSIVITY o Description Hydrodynamic dispersion is the macroscopic spreading of a dissolved constituent plume due to effects of chemical diffusion and mechanical dispersion Mechanical dispersion is caused by the small scale variations in flow velocity through porous media causing the paths of solutes to spread from the overall direction of groundwater flow Transverse vertical hydrodynamic dispersivity defines how strong the mechanical mixing component is For the Dandy Sale model this value is used to define the vertical distribution of concentration at the vertical plane source see equation 3 of Appendix A 2 1 and spreads the plume out vertically as it progresses downstream F
45. x 10 Dichloroethene cis1 2 1 38 8 00 x 107 MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 69 Organic Carbon Petitioning CHEMICAL PARAMETER DATABASE Coefficient CS Chemical Name log Koc 20 25 C LE log 1 kg ar Dichloroethene trans1 2 1 46 1 75 x 10 Ethylbenzene 1 98 6 00 x 107 Ethylene glycol 0 90 1 00 x 10 Fluoranthene 4 58 2 06 x 10 Fluorene 3 86 1 69 x 10 Hexane n 2 68 1 30 x 10 Indeno 1 2 3 c d Pyrene 7 53 7 17 x 10 Methanol 0 69 1 00 x 10 Methylene chloride 1 23 1 54 x 10 Methyl ethyl ketone 0 28 2 18 x 10 Methyl t Butyl Ether 1 08 4 80 x 10 Naphthalene 3 11 3 29 x 10 Phenanthrene 4 15 1 60 x 10 Phenol 144 9 30 x 10 Pyrene 4 58 1 60 x 10 Tetrachloroethane 1 1 2 2 0 00 7 18 x 10 Tetrachloroethene 2 43 1 43 x 107 Toluene 2 13 5 15 x 107 Trichlorobenzene 3 91 3 03 x 10 Trichloroethane 1 1 1 2 45 1 26 x 10 Trichloroethane 1 1 2 1 75 5 93 x 10 Trichloroethene 1 26 1 00 x 10 Trichlorofluoromethane 2 49 2 47 x 10 Vinyl Chloride 0 39 2 54 x 107 Xylene mixed isomers 2 38 1 98 x 10 Xylene m 3 20 1 58 x 10 Xylene o 2 11 1 75 x 107 Values obtained from Natural Attenuation of Fuels and Chlorinated Solvents in the Subsurface by Wiedemeier et al 1999 Appendix B MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 70 GEOLOGIC PARAMETER DATABASE Parameter Value Units Hydraulic Conductivity Clean sands 0 001 1 cm s Clays 1x 10 cm s Gravels gt 1 cm s
46. year 1952 Site history Source removed in year 1996 Site history Field Data for Comparison TCE Concentration in MW 01 transmissive zone Conc ug L Tm 2 371 3 162 1 957 1 000 1 468 908 Groundwater sampling MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 116 CASE STUDY 1A INDUSTRIAL SITE CONNECTICUT SQUARE ROOT MODEL SRM Summary e The Toolkit SRM was used to estimate TCE groundwater concentrations in the transmissive zone following DNAPL remediation at an industrial facility Uncertainties associated with the estimates were also evaluated Toolkit input parameters are shown on Figure 1 2 e The surficial sandy aquifer at the site is underlain by a thick silty aquitard Heavy historical industrial pumping resulted in a long term downward hydraulic gradient across the aquitard Chapman and Parker 2005 e SRM Plume Analysis model Section 2 was used to estimate the groundwater concentrations e Site hydrogeological data was entered in Section 3 transport parameters in Section 4 plume characteristics in Section 5 source loading information in Section 6 and field data for comparison in Section 7 e Site specific values as documented by Chapman and Parker 2005 were available for all parameters except molecular diffusion coefficient in free water and apparent tortuosity factor exponent For these Toolkit default values were used e Since exact source concentrations
47. 0 CASE STUDY 3 FORMER DRY CLEANER FLORIDA DANDY SALE MODEL PCE Soil Concentration in Low k Zone mg kg 5 10 15 20 25 3 gO 800000 K 0 5 5 j Observed PCE Not calibrated Calibrated Figure 3 3 Comparison of Toolkit Simulated and Field Observations OU3 3 65 ft from Source Note the red line did not calibrate well due to the low seepage velocity estimate for the site see text This point was then calibrated using a larger seepage velocity The match between actual and modeled results is very close c 2 x 1 s o o 4 lt 4 a GU D MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 151 MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 152
48. 3E 01 L kg Year format yyyy Lateral Distance from Source x m Depth into Low k Zone Z m T Gas EE Show Graph Figure 1 7 DSM Input Parameters Plume Zone Evaluation Initial MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 127 CASE STUDY 1B INDUSTRIAL SITE CONNECTICUT DANDY SALE MODEL Low k Zone Aqueous Concentration mg L at x 280 m for Year 2000 Concentration in Low k Zone mg L 0 00E 00 1 00E 01 2 00E 01 3 00E 01 4 00E 01 5 00E 01 6 00E 01 ENEE SE 1 1 E o c o N X o be o 2 IE I ZS a o a 3 0 Restore Original Graph See Results for Lateral Distance from Source x 280 Low Permeability Zone See 2 D Low k Aqueous Conc See Trans Zone Aqueous Conc Export Low k 2 D Data HELP See Low k Aq Conc vs Dist Aqueous Mass ETS ka Next Step a See Trans Zone Mass Discharge SE i Sorbed Mass EET o Save Data m Data input See Trans Zone Sorbed Conc Total Mass Aay ko See 2 D Low k Sorbed Conc See Trans Zone Total Conc Return to Main Screen See 2 D Low Perm Total Conc Figure 1 8 DSM Output Plume Area Low k Zone Concentrations Initial MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 128 CASE STUDY 1B DSM Data Input Screen Matrix Diffusion Toolkit INDUSTRIAL SITE DANDY SALE MODEL Version 1 0 Site Location and ID Industrial Site Connecticut SYSTEM UNITS SI Units
49. 4 Low k Aqueous Mass Purpose Determine the low k aqueous phase mass output in the Dandy Sale Model of the Matrix Diffusion Toolkit Given The vertical plane source produces a plume in the transmissive zone that loads up the low k zone due to diffusion This vertical plane source is shut off and diffusion results in a release of contaminants from the low k zone Assumptions The model makes the following assumptions 1 A source considered to be a thin pool is introduced at the contact between the two layers upgradient of x 0 2 A loading period occurs where there is a constant concentration of contaminants in the transmissive zone that drives contaminants into the low k zone 3 A release period occurs where the transmissive zone is assumed to have no concentration and an upper range estimate of release from the low k zone is generated 4 There is no degradation in either layer 5 Both layers are uniform homogeneous isotropic and infinite in the z direction perpendicular to groundwater flow 6 One dimensional 1 D advective transport in the transmissive layer parallel to the boundary of the layers is accompanied by transverse dispersion and diffusion 7 There is no longitudinal dispersion in the transmissive layer 8 1 D transverse diffusion transport occurs in the low k layer 9 Retardation of contaminants in both layers is based on instantaneous equilibrium between aqueous and sorbed phases Summary Activ
50. 46 Sand fine 0 26 0 53 Silt 0 34 0 61 Clay 0 34 0 60 SEDIMENTARY ROCKS Sandstone 0 05 0 30 Siltstone 0 21 0 41 Shale 0 0 10 CRYSTALLINE ROCKS Dense crystalline rocks 0 0 05 Koerner 1984 reports these values for unit weight for saturated soils note no dry bulk density values are reported for these materials Glacial till very mixed grain 0 20 Soft glacial clay 0 57 Stiff glacial clay 0 37 Soft slightly organic clay 0 66 Soft very organic clay 0 75 Soft bentonite 0 84 One fractured microcrystalline limestone in Virginia had matrix porosities ranging from 0 0004 to 0 0065 GSI Environmental Source of Data Typically estimated Occasionally obtained through physical property testing of site soil samples How to Enter Data Enter directly Note that if the low k zone description is selected from the drop down list the Toolkit provides a default value for the parameter MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 47 DSM DATA ENTRY PARAMETER TRANSMISSIVE ZONE SEEPAGE VELOCITY V cm sec ft or m day ft or m yr Description Actual interstitial groundwater velocity equaling Darcy velocity divided by effective porosity Typical Values 1 1500 ft yr 0 3 457 m yr Source of Data Calculated by multiplying hydraulic conductivity by hydraulic gradient and dividing by effective porosity V K x i ne Itis strongly recommended that actual site data be used for hydraulic conductivity and hydra
51. 5 CASE STUDY 2A SAND TANK STUDY SQUARE ROOT MODEL fluorescein comparison of simulated and observed concentrations was obtained by using the maximum observed concentration as the source concentration For bromide concentrations were underestimated closer to the source cut off and overestimated towards the end of the simulations Although there are various combinations of input parameters could be varied to improve the comparison of simulated and observed concentrations for this analysis the parameter adjusted was the loading concentration Increasing the loading concentration yielded a close match to the observed fluorescein concentration vs time data These results show that the Square Root Model can simulate complex heterogeneous systems that don t meet all of the assumptions and still provide useful simulation results that are within an order of magnitude References Chapman S W and B L Parker T C Sale and L A Doner 2012 Testing high resolution numerical models for analysis of contaminant storage and release from low permeability zones J Cont Hydrology 136 137 106 116 MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 136 CASE STUDY 2A SAND TANK STUDY SQUARE ROOT MODEL SRM Data Input Screen DATA INPUT INSTRUCTIONS Matrix Diffusion Toolkit Version 1 0 LLL Enter value directly Wl value calculated by Toolkit Do not enter data Site Location and ID Sand Tank stud 1 SYSTEM UNITS 2 ANALYSIS TYPE 5 PLU
52. 96 Source of Data Pump tests or slug tests at the site It is strongly recommended that actual site data be used for all matrix diffusion evaluations How to Enter Data 1 Select units and 2 Enter directly PARAMETER HYDRAULIC GRADIENT i Units ft ft or m m Description The slope of the potentiometric surface In unconfined aquifers this is equivalent to the slope of the water table Typical Values 0 0001 0 1 ft ft 0 0001 0 1 m m Source of Data Calculated by constructing potentiometric surface maps using static water level data from monitoring wells and estimating the slope of the potentiometric surface MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 63 DISSOLUTION MODEL DATA ENTRY Step 3 Transport PARAMETER Description How to Enter Data PARAMETER Units Description Typical Values Source of Data How to Enter Data PARAMETER Units Description Typical Values Source of Data How to Enter Data KEY CONSTITUENT Constituent of interest Enter directly or choose from drop down list VERTICAL TRANSVERSE DISPERSIVITY av Dispersion refers to the process whereby a dissolved solvent will be spatially distributed because of mechanical mixing and chemical diffusion in the aquifer These processes develop the plume shape that is the spatial distribution of the dissolved solvent mass in the aquifer Selection of dispersivity values is a difficult process given the impracticability of
53. Box 7 10E 01 Concentration ug L Approximate Width Width of Black Box 3 00E 02 Mass Discharge g day Highest Historical Concentration in Black Box 1 88E 00 Mass kg 1 88E 00 Concentration of Contour Line in Black Box 1 88E 00 re Paste Example Representative Concentration OK to Override Next Highest Concentration Zone Blue Box in Picture See eier Save paa imate Length Length of Blue B 7 10E 01 ow Gra pana Valse o sat P Return to Model Selection Screen Return to Main Screen Approximate Width Width of Blue Box 3 00E 02 Figure 2 3 SRM Input Parameters Fluorescein Calibrated Source concentration changed to 1 88 mg L MATRIX DIFFUSION TOOLKIT v USER S MANUAL Y 138 CASE STUDY 2A SAND TANK STUDY SQUARE ROOT MODEL Time day 60 A Observed Fluorescein Not Calibrated Calibrated n zl BD E 2 o ke re Kl o o M Figure 2 4 Comparison of SRM Green Lines against Observed Concentrations Fluorescein The dark green line represents output using initial parameters The light green line represents the calibrated model output The overall shape of the uncalibrated model result matched the data and was within one order of magnitude The calibrated model was a very close match to experimental data MATRIX DIFFUSION TOOLKIT v USER S MANUAL Y 139 CASE STUDY 2A SAND TANK STUDY SQUARE ROOT MODEL SRM Data Input Screen DATA INPUT INSTR
54. DATA Description Loads data files saved through the Toolkit DO NOT EDIT ANY TOOLKIT FILES Editing files may cause the Toolkit to crash PARAMETER RETURN TO MODEL SELECTION SCREEN Returns to the Model Selection Screen PARAMETER RETURN TO MAIN SCREEN Returns to the Matrix Diffusion Toolkit Main Screen MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 68 CHEMICAL PARAMETER DATABASE Organic Carbon Petitioning Coefficient Seley Chemical Name log Koc 20 25 C 20 25 E log 1 kg WE Acetone 0 24 1 00 x 10 Acenaphthene 3 85 3 93 x 10 Acenaphthylene 4 00 3 93 x 10 Anthracene 4 15 4 50 x 10 Benzene 1 58 1 75 x 10 Benzoic acid 1 83 6 22 x 10 Benzo a Anthracene 6 14 5 70 x 10 Benzo b Fluoranthane 5 74 1 47 x 10 Benzo k Fluoranthene 5 74 4 30 x 10 Benzo g h i Perylene 6 20 7 00 x 107 Benzo a Pyrene 5 59 1 20 x 10 Bromodichloromethane 1 85 6 22 x 10 Butanol n 0 74 7 70 x 10 Carbon disulfide 2 47 2 30 x 10 Carbon tetrachloride 2 67 7 62 x 10 Chlorobenzene 2 46 4 45 x 10 Chloroethane 1 25 2 00 x 10 Chloroform 1 93 9 64 x 10 Chloromethane 1 40 4 00 x 10 Chlorophenol 2 2 11 2 85 x 10 Chrysene 5 30 1 80 x 10 Dibenzo a h Anthracene 5 87 5 00 x 107 Dibromochloromethane 2 05 5 25 x 10 Dichlorobenzene 1 2 o 3 32 1 50 x 10 Dichlorobenzene 1 4 p 3 33 1 45 x 10 Dichlorodifluoromethane 2 12 1 98 x 10 Dichloroethane 1 1 1 76 5 00 x 10 Dichloroethane 1 2 1 76 8 69
55. E z Uis PRM PRM Vertical depth of the low k zone from the source for displaying matrix diffusion results How to Enter Data to Enter Data Enter Enter directly 00 PARAMETER NEXT STEP SHOW GRAPH Proceeds to the results of matrix diffusion analysis PARAMETER SHOW PREVIOUS DATA Shows the output for previously run analysis PARAMETER NEW SITE CLEAR DATA MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 56 DSM DATA ENTRY Description Clears ALL data related to the DSM model in the Toolkit memory banks Use this button to start a new project PARAMETER PASTE EXAMPLE Description Clears ALL data related to the DSM model in the Toolkit memory banks and pastes an example dataset The example dataset used in the Toolkit is obtained from Chapman and Parker 2005 PARAMETER SAVE DATA Description Saves all the DSM model data DO NOT ADD ANY EXTENSIONS TO FILE NAME WHEN SAVING Note that this option does not save any edits performed on the graphs by the user To save such edits use the save function of Excel and save the entire Toolkit file PARAMETER LOAD DATA Description Loads data files saved through the Toolkit DO NOT EDIT ANY TOOLKIT FILES Editing files may cause the Toolkit to crash PARAMETER RETURN TO MODEL SELECTION SCREEN Returns to the Model Selection Screen PARAMETER RETURN TO MAIN SCREEN Returns to the Matrix Diffusion Toolkit Main Screen MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 57
56. E 78 APPENDICES rtr eerie elected Ee 79 APPENDIX A 1 SQUARE ROOT MODEL see EEN EEN 80 Appendix A 1 1 Estimation of Mass Discharge rnnt 80 Appendix A 1 2 Estimation of Concentration in Transmissive Zone eee ee eee eee e 81 Appendix A 1 3 Estimation of Mass in Transmissive Zone 82 APPENDIX A 2 DANDY SALE MODEL ss ssssseee sees sese nner eenn enen 84 Appendix A 2 1 Low k Aqueous Concentration ss ses eee eee eee 84 Appendix A 2 2 Low k Sorbed Concentration sse 88 Appendix A 2 3 Low k Total CGoncentraton seen 90 Appendix A 2 4 Low k Aqueous Mass 92 Appendix A 2 5 Low k Sorbed Mass eene nennen 94 Appendix A 2 6 Low k Total Mass seen tenen nnne 96 Appendix A 2 7 Transmissive Layer Aqueous Concentration eee eee 98 Appendix A 2 8 Transmissive Layer Sorbed Concentration eee eee eee eee 105 Appendix A 2 9 Transmissive Layer Total Concentration sss 106 Appendix A 2 10 Transmissive Layer Aqueous Mass 108 Appendix A 2 11 Transmissive Layer Sorbed Mass 109 Appendix A 2 12 Transmissive Layer Total Mass enne 110 APPENDIX A 3 PROBABILITY DISTRIBUTIONS s sssess eee ss esser eee eenn eenn nenen nner 111 Appendix A 3 1 Normal Distributions sss sese ee eee eee 111 Appendix A 3 2 Lognormal Distributions 111 Appendix A 3 3 Uniform Distributions sss sese eee e xe eee 112 CASE STUDIES eerie Bede S 113 CASE STUDY 1 INDUSTRIAL SITE CONNEC TIC UT sss sese essere eenn eenn
57. ES AND LIMITATIONS o A release period where the transmissive zone is assumed to have no concentration and an upper range estimate of release out of the low k zone is generated o That is the system is assumed to be of the on off type with a defined loading period that extends for a certain period of time that then switches to a release period where any concentration in the transmissive zone that originates from non back diffusion sources is instantly switched off e Assumes an infinitely thick low k zone which in practice means the low k zone is at least 1 meter thick for sites where matrix diffusion has been occurring for several decades Thinner low k zones can be modeled but with more uncertainty in the final results Case Studies 2A and 2B show both models in the Toolkit being applied to a tank study with layers as thin as 0 03 meters where the theoretical penetration in the low k zones during the 124 day test period was about 0 25 meters Despite not corresponding to the assumption of a low k zone that is thicker than the penetration depth the model outputs were within an order of magnitude of the actual measured concentrations from the tank study e Assumes no degradation in the low k zone e Torun the Monte Carlo analysis users need to estimate what type of statistical distribution best fits the input data and what values best describe the distribution In many cases data will be unavailable to make these estimates s
58. Enclosure Figure 1 1 Site Location Map Based on Chapman and Parker 2005 Figure 1 MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 115 CASE STUDY 1A INDUSTRIAL SITE CONNECTICUT SQUARE ROOT MODEL A Square Root Model SRM Input Data Data Type Source of Data Hydrogeology Low k zone description Low k zone porosity Darcy velocity Transport Low k Zone Key constituent Molecular diffusion coefficient in free water Apparent tortuosity factor exponent Retardation factor Plume Characteristics High concentration zone Approximate length Approximate width Highest concentration in black box Concentration of contour line in black box Representative concentration Next highest concentration zone Approximate length Approximate width Concentration of contour line in blue box Representative concentration Uncertainty in plume concentration estimations silt 0 43 0 13 m d TCE 9 1E 10 m sec Boring logs Estimated Calculated Site history Literature Toolkit default 0 33 Literature Toolkit default 1 2 Calculated using measured faction organic carbon Based on area of affected 330 m groundwater plume 300 m 37 000 ug L 37 000 ug L 37 000 ug L Same as black box 330 m 300 m 37 000 ug L 37 000 ug L 10 factor of General Source loading starts in
59. INE IN BLUE BOX Description Standard Method Leave this blank and just enter the historical loading concentration for the first modeled area in Representative Concentrations Contour Map Method Concentration of contour line denoted by the blue box in the Toolkit input screen figure Typical Values 0 0001 1 000 mg L Source of Data Standard Method this is not needed Contour Map Method Use a contour map from the highest concentration period where groundwater samples were collected For example if concentrations have been decreasing use a concentration contour map from 1990 and not 2012 How to Enter Data Enter directly PARAMETER REPRESENTATIVE CONCENTRATION C 2 Description Representative historical loading concentration of second modeled area denoted by the blue box in the Toolkit input screen figure This value is a key parameter that can be changed during the calibration process to increase or decrease the simulated mass discharge concentration or mass to better match field data see the beginning of MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 33 SRM DATA ENTRY 1 this section S Typical Values 0 20 000 mg L Source of Data Standard Method This could be the same sources of data used for Cs1 described earlier but for a second part of the modeled area Information from either Data Source 1 or Data Source 2 can be used Contour Map Method This is calculated automatically by the model from the previ
60. ME CHARACTERISTICS CONT D SI Units O English Units C Source Zone Analysis Plume Analysis C PRB Analysis B Concentration of Contour Line in Blue Box 1 00E 00 Representative Concentration OK to Override ug L Restore B Uncertainty in Plume Concentration Estimations 3 HYDROGEOLOGY Low k Zone Description Low k Zone Porosity Transmissive Zone Darcy Velocity me Calculate va 4 TRANSPORT LOW k ZONE 6 GENERAL Key Constituent User Input z Source Loading Starts in Year 2006 format yyyy Molecular Diffusion Coefficient in Free Water Source Removed in Year format yyyy Apparent Tortuosity Factor Exponent T o Retardation Factor CalculateR 5 PLUME CHARACTERISTICS See Back Diffusion Results from Year 2006 format yyyy to Year 2006 36 format yyyy in Intervals of 0 00274 yrs 7 FIELD DATA FOR COMPARISON High Concentration Zone Black Box in Picture Year Approximate Length Length of Black Box 7 10E 01 Concentration ug L Approximate Width Width of Black Box 3 00E 02 Mass Discharge g day Highest Historical Concentration in Black Box 1 00E 00 Mass kg 1 00E 00 Concentration of Contour Line in Black Box 1 00E 00 re Paste Example Representative Concentration OK to Override Next Highest Concentration Zone Blue Box in Picture See eier Save paa imate Length Length of Blue B 7 10E 01 ow Gra pene CET P Return to Model Selection Screen Ret
61. Matrix Diffusion Toolkit USER S MANUAL Version 1 0 September 2012 Groundwater Flow Direction a a Transmissive Former Zone Source Loading WY q N S NN gt N Diffusion from low k zone causing mass discharge into transmissive zone S K Farhat C J Newell M A Seyedabbasi J M McDade N T Mahler GSI ENVIRONMENTAL INC HOUSTON TEXAS T C Sale DS Dandy J J Wahlberg COLORADO STATE UNIVERSITY FORT COLLINS COLORADO ESTCP ESTCP DISCLAIMER The Matrix Diffusion Toolkit is available as is Considerable care has been exercised in preparing this manual and software product however no party including without limitation the United States Government GSI Environmental Inc Colorado State University the authors and reviewers make any representation or warranty regarding the accuracy correctness or completeness of the information contained herein and no such party shall be liable for any direct indirect consequential incidental or other damages resulting from the use of this product or the information contained herein Information in this publication is subject to change without notice Implementation of the Matrix Diffusion Toolkit and interpretation of the predictions of the models are the sole responsibility of the user CITE USING Farhat S K C J Newell T C Sale D S Dandy J J Wahlberg M A Seyedabbasi J M McDade and N T Mahler 2012 Matrix Dif
62. OOLKIT Y USER S MANUAL Y 89 APPENDIX A 2 DANDY SALE MODEL Appendix A 2 3 Low k Total Concentration Purpose Determine the low k total concentration output in the Dandy Sale Model of the Matrix Diffusion Toolkit Given The vertical plane source produces a plume in the transmissive zone that loads up the low k zone due to diffusion This vertical plane source is shut off and diffusion results in a release of contaminants from the low k zone Assumptions The model makes the following assumptions 1 A source considered to be a thin pool is introduced at the contact between the two layers upgradient of x 0 2 A loading period occurs where there is a constant concentration of contaminants in the transmissive zone that drives contaminants into the low k zone 3 A release period occurs where the transmissive zone is assumed to have no concentration and an upper range estimate of release from the low k zone is generated 4 There is no degradation in either layer 5 Both layers are uniform homogeneous isotropic and infinite in the z direction perpendicular to groundwater flow 6 One dimensional 1 D advective transport in the transmissive layer parallel to the boundary of the layers is accompanied by transverse dispersion and diffusion 7 There is no longitudinal dispersion in the transmissive layer 8 1 D transverse diffusion transport occurs in the low k layer 9 Retardation of contaminants in both layers is bas
63. R S MANUAL Y 62 DISSOLUTION MODEL DATA ENTRY PARAMETER SEEPAGE VELOCITY V cm sec ft or m day ft or m yr Description Actual interstitial groundwater velocity equaling Darcy velocity divided by effective porosity Typical Values 1 1500 ft yr 0 3 457 m yr Source of Data Calculated by multiplying hydraulic conductivity by hydraulic gradient and dividing by effective porosity V K x i ne Itis strongly recommended that actual site data be used for hydraulic conductivity and hydraulic gradient data parameters effective porosity can be estimated How to Enter Data 1 Select units and enter directly or 2 Calculate entering values for a Hydraulic conductivity b Hydraulic gradient and c Pressing the Calculate V button PARAMETER HYDRAULIC CONDUCTIVITY K cm sec ft or m day ft or m yr Description Measure of the permeability of the transmissive layer To characterize concentrations in the transmissive layer representative measurements are required for the Darcy velocity or both the hydraulic flow gradient and the hydraulic conductivity of the flow system Representative measurements of hydraulic conductivity of the transmissive layer should be obtained at one or more locations using appropriate slug test or pumping test methods Newell et al 2003 Typical Values Silts 1x108 1x10 3 cm s Silty sands 1x10 1x1071 cm s Clean sands 1x10 3 1 cm s Gravels gt 1 cm s Newell et al 19
64. S Uncertainty Analysis Perform Uncertainty Analysis Uncertainty in parameter estimates is a key issue in estimating matrix diffusion effects The Toolkit provides two options for analyzing this uncertainty One option performed automatically provides a lower range mostly likely value and an upper range for estimated outputs based on the specified source area concentrations The second option Advanced Uncertainty Analysis utilizes a Monte Carlo type approach to analyze uncertainty in the actual concentration porosity apparent tortuosity factor exponent and retardation factor measurements With this tool groundwater practitioners can estimate the accuracy of the hydrologic measurements that are being used for the matrix diffusion calculation PARAMETER ADVANCED UNCERTAINTY ANALYSIS EVALUATE HOW UNCERTAINTY IN INPUT DATA AFFECTS TOTAL MASS FLUX Description This module uses the Monte Carlo approach to analyze uncertainty in the actual concentration porosity apparent tortuosity factor exponent and retardation factor measurements In the Monte Carlo type approach a random number is generated for every value of concentration porosity apparent tortuosity factor exponent and retardation factor entered by the user This set of random inputs is then used to calculate mass discharge in both the low k and transmissive zones concentration in the transmissive zone and mass in the transmissive zone Repeating this procedure a large num
65. S Unit system to perform matrix diffusion calculations in SI System meters etc or English Units feet etc How to Enter Data Choose the appropriate radio button Step 2 Hydrogeology Description Description of the transmissive zone Sand gravel and silt are provided as the three selections How to Enter Data Choose from drop down list or enter directly PARAMETER TRANSMISSIVE ZONE EFFECTIVE POROSITY ne Description Dimensionless ratio of the volume of voids to the bulk volume of the surface soil column matrix Note that total porosity is the ratio of all voids including non connected voids to the bulk volume of the aquifer matrix Differences between total and effective porosity reflect lithologic controls on pore structure In unconsolidated sediments coarser than silt size effective porosity can be less than total porosity by 2 5 e g 0 28 vs 0 30 Smith and Wheatcraft 1993 Typical Values Toolkit default values provided are averages of the ranges below Gravel 0 10 0 35 Coarse Sand 0 20 0 35 Fine Sand 0 10 0 30 Medium Sand 0 15 0 30 From Wiedemeier et al 1999 originally from Domenico and Schwartz 1990 and Walton 1988 Source of Data Typically estimated Occasionally obtained through physical property testing of site soil samples One commonly used value for silts and sands is 0 25 The ASTM RBCA Standard ASTM 1995 includes a default value of 0 38 to be used primarily for unconsolidated
66. SER S MANUAL Y 82 APPENDIX A 1 SQUARE ROOT MODEL M tt 24C e Ve i r MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 83 APPENDIX A 2 DANDY SALE MODEL Appendix A 21 Low k Aqueous Concentration Purpose Determine the low k aqueous concentration output in the Dandy Sale Model of the Matrix Diffusion Toolkit Note This derivation was originally developed by Sale et al 2008b and described in detail in Sale et al 2008b and Bolhari 2012 Given There is source material in a transmissive zone that loads up a downgradient low k zone during the loading period before the source is removed Assumptions The Toolkit uses a simplified conceptual model of a two layer aquifer system a transmissive layer above a low k layer Figure A 2 1 1 A source e g DNAPL is introduced at the contact between the two layers As shown in the figure x is in the direction of groundwater flow and z the direction perpendicular to groundwater flow The edge of the source at the interface between the two layers is designated x 0 and z 0 with both x and z increasing with distance away from the source edge In this model z is designated as the vertical depth from the source in the low k layer and z the height in the transmissive zone Transmissive Zone Source 4 lt Low k Zone Results Calculated Here Figure A 2 1 1 The two layer scenario conceptual model Top Active Source Bottom Depleted Source The model makes the f
67. SION TOOLKIT v USER S MANUAL Y 117 CASE STUDY 1A INDUSTRIAL SITE CONNECTICUT SQUARE ROOT MODEL within an order of magnitude to represent this difficult to model complex process is a significant improvement and consequently this model provides very useful information Based on Toolkit SRM modeling more than 500 years will be required for the plume to reach an MCL of 5 ug L This compares well to Chapman and Parker s more sophisticated modeling that indicated concentrations will remain much above the MCL for centuries MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 118 CASE STUDY 1A INDUSTRIAL SITE CONNECTICUT SQUARE ROOT MODEL SRM Data Input Screen j DATA INPUT INSTRUCTIONS Matrix Diffusion Toollat Version 1 0 Enter value directly Value calculated by Toolkit Do not enter data Site Location and ID Industrial Site Connecticut th SYSTEM UNITS 2 ANALYSIS TYPE 5 PLUME CHARACTERISTICS CONT D SI Units English Units Source Zone Analysis Plume Analysis PRB Analysis E Concentration of Contour Line in Blue Box 3 70E 04 ug L 3 HYDROGEOLOGY Representative Concentration OK to Override C 2 3 T0E 04 ug L Restore Low k Zone Description Silt sit EN Uncertainty in Plume Concentration Estimations factor of 10 Low k Zone Total Porosity 0 431 Transmissive Zone Darcy Velocity 0 13 ma Calculate Vd 4 TRANSPORT Low k Zone 6 GENERAL Key Constituent TCE TCE Source Loading
68. STUDY 1B INDUSTRIAL SITE CONNECTICUT DANDY SALE MODEL Low k Zone Aqueous Concentration mg L at x 280 m for Year 2000 Concentration in Low k Zone mg L 0 00E 00 5 00E 00 1 00E 01 1 50E 01 2 00E 01 2 50E 01 3 00E 01 4 00E 01 0 0 i 4 E o G N St 3 4 2 Restore Original Graph See Results for Lateral Distance from Source x 280 Im ili e Low Permeability Zone AMEDEO MER SS See Trans Zone Aqueous Conc Export Low k 2 D Data See Low k Aq Conc vs Dist Aqueous Mass 5 4E 02 ko Next Step e Ed 3 See Trans Zone Mass Discharge teen Sorbed Mass EX ko 2 SeT Z caned Retu Total Mass Er ko Save Data e 5ea 2D low Sorbed Conc ee Trans Zone Sor onc See Trans Zone Total Conc Return to Main Screen See 2 D Low Perm Total Conc Figure 1 11 DSM Output Plume Area Low k Zone Concentrations Calibrated MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 131 CASE STUDY 1B INDUSTRIAL SITE CONNECTICUT DANDY SALE MODEL Using C 475 mg L ML 10 Aug 2000 Using C 1100 mg L 3 o E o N x g z o el E Fei a eo a 40 60 TCE Concentration mg L Figure 1 12 Comparison of DSM Plume Area Low k Concentrations Red and Blue Lines with Observed Concentrations in ML 10 in 2000 Based on Figure 6b of Chapman and Parker 2005 The calibrated value C 1100 mg L loading concentration for first 26 years resulted in a better match to actual site da
69. Silts 1x 10 4 x 10 cm s Silty sands 15010205610 cm s Total Porosity Basalt 0 03 0 35 Clay 0 34 0 60 Coarse Gravel 0 24 0 36 Fine Gravel 0 25 0 38 Fine Sand 0 26 0 53 Coarse Sand 0 31 0 46 Limestone 0 0 0 5 Sandstone 0 05 0 30 Shale 0 0 0 10 is Silt 0 34 0 61 Siltstone 0 21 0 41 C Effective Porosity Clay 0 01 0 20 Fine Gravel 0 2 0 35 C Medium Gravel 0 15 0 25 Coarse Gravel 0 1 0 25 Sandy Clay 0 03 0 2 E Loess 0 15 0 35 Peat 0 3 0 5 C Silt 0 01 0 3 Gravely Sand 0 2 0 35 Fine Sand 0 10 0 30 Medium Sand 0 15 0 30 5 Coarse Sand 0 20 0 35 MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 71 GEOLOGIC PARAMETER DATABASE Parameter Value Units Effective Porosity Glacial Sediments 0 05 0 2 Limestone 0 01 0 24 Unfractured Limestone 0 001 0 05 Sandstone 0 1 0 4 Siltstone 0 01 0 35 Fractured Granite 0 00005 0 01 Volcanic Tuff 0 02 0 35 Dry Bulk Density Clay 1 00 2 40 g cm Silt lt g cm Granite 2 24 2 46 g cm Fine Sand 1 37 1 81 g cm Medium Sand 1 37 1 81 g cm Coarse Sand 1 37 1 81 g cm Sandstone 1 60 2 68 g cm Gravel 1 36 2 19 g cm Limestone 1 74 2 79 g cm Notes From Newell et al 1996 2 From Wiedemeier et al 1995 3 From Wiedemeier et al 1999 originally from Domenico and Schwartz 1990 and W
70. Starts in Year format yyyy Molecular Diffusion Coefficient in Free Water 9 10E 10 m2 sec E Source Removed in Year format yyyy Apparent Tortuosity Factor Exponent 3 30E 01 Retardation Factor 1 201 _ Calculate R 2 5 PLUME CHARACTERISTICS See Release Period Results from Year format yyyy to Year format yyyy in Intervals of yrs 7 FIELD DATA FOR COMPARISON High Concentration Zone Black Box in Picture Year 1998 1999 2001 Approximate Length Length of Black Box 3 30E 02 m Concentration ug L 3832 2371 1957 Approximate Width Width of Black Box 3 00E 02 m Mass Discharge g day Highest Historical Concentration in Black Box 3 70E 04 ug E Mass kg Concentration of Contour Line in Black Box 3 70E 04 ug L Representative Concentration OK to Override 3 70E 04 ug L Restore Next Highest Concentration Zone Blue Box in Picture Next Step HEC E ER GE Approximate Width Width of Blue Box 3 00E 02 Figure 1 2 SRM Input Parameters MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 119 CASE STUDY 1A INDUSTRIAL SITE CONNECTICUT SQUARE ROOT MODEL e er n Se CERERI DECH mti eet ESCH PENES s EE Weg ET ceca ce ee jc Note Negative mass discharge values represent diffusion into the low k zone from the transmissive zone Positive values represent diffusion from the low k zone into the
71. UCTIONS Matrix Diffusion Toolkit Version 1 0 LLL Enter value directly Wl value calculated by Toolkit Do not enter data Site Location and ID Sand Tank stud 1 SYSTEM UNITS 2 ANALYSIS TYPE 5 PLUME CHARACTERISTICS CONT D SI Units O English Units C Source Zone Analysis Plume Analysis C PRB Analysis B Concentration of Contour Line in Blue Box 1 00E 00 Representative Concentration OK to Override ug L Restore B Uncertainty in Plume Concentration Estimations 3 HYDROGEOLOGY Low k Zone Description Low k Zone Porosity Transmissive Zone Darcy Velocity me e Calculate va 4 TRANSPORT LOW k ZONE 6 GENERAL Key Constituent User Input z Source Loading Starts in Year 2006 format yyyy Molecular Diffusion Coefficient in Free Water Source Removed in Year format yyyy Apparent Tortuosity Factor Exponent o Retardation Factor CalculateR 5 PLUME CHARACTERISTICS See Back Diffusion Results from Year 2006 format yyyy to Year 2006 36 format yyyy in Intervals of 0 00274 yrs 7 FIELD DATA FOR COMPARISON High Concentration Zone Black Box in Picture Year Approximate Length Length of Black Box 7 10E 01 Concentration ug L Approximate Width Width of Black Box 3 00E 02 Mass Discharge g day Highest Historical Concentration in Black Box 1 00E 00 Mass kg 1 00E 00 Concentration of Contour Line in Black Box 1 00E 00 re Paste Example Representative Concentr
72. after all of the DNAPL is gone Because of the simplifying assumptions discussed above results provided by the Toolkit will be planning level information But these results can help you think about these different questions and tell you what might happen MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 8 FREQUENTLY ASKED QUESTIONS What if don t have a two layer system at my site Can I still use the Matrix Diffusion Toolkit Yes with some limitations If you have multiple thick low k units within a transmissive zone you can determine the number of layers and multiply the model outputs by that number see Inset 1 on the next page Case Study 2 shows an example of a four layer system with eight interfaces two for each layer that was modeled successfully with the Toolkit Because both models in the Toolkit assume a single layer two interface problem the final concentration results from the Toolkit were multiplied by eight to get an estimate of the effect of all eight interfaces on concentration The end results were concentrations that matched measured concentration output from the tank study to within an order of magnitude Do need sampling data from the low k zones to run the Toolkit models If you want results with a high level of confidence then calibrating the Toolkit output to soil concentration data you collect from the low k zones would be very important However if you want to learn more about the potential impacts of matr
73. alton 1988 MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 72 REFERENCES AFCEE 2007 Source Zone Initiative Final Report Submitted to Air Force Center for Environmental Excellence May 2007 Adamson D T 2012 GSI Environmental Inc Houston Texas Personal communication American Society for Testing and Materials 1995 Standard Guide for Risk Based Corrective Action Applied at Petroleum Release Sites ASTM E 1739 95 Philadelphia Pennsylvania Aziz C E C J Newell J R Gonzales P E Haas T P Clement and Y Sun 2000 BIOCHLOR Natural Attenuation Decision Support System User s Manual Version 1 0 U S EPA Office of Research and Development EPA 600 R 00 008 Washington D C January 2000 www gsi net com Bergin M S and J B Milford 2000 Application of Bayesian Monte Carlo analysis to a Lagrangian photochemical air quality model Atmospheric Environment 34 781 792 Bird R B W E Stewart and E N Lightfoot 1960 Transport Phenomena John Wiley and Sons Inc Bolhari A 2012 Feasibility of Treating Chlorinated Solvents Stored in Low Permeability Zones in Sandy Aquifers PhD Dissertation to be completed Colorado State University Fort Collins Colorado Chapman S W and B L Parker 2005 Plume persistence due to aquitard back diffusion following dense nonaqueous phase liquid source removal or isolation Water Resources Research 41 W12411 doi 10 1029 2005WR004224 Charbeneau R J
74. ameters is strongly recommended How to Enter Data 1 Select units and enter directly or 2 Calculate by pressing the Calculate Vd button and entering values for a Hydraulic conductivity and b Hydraulic gradient PARAMETER TRANSMISSIVE ZONE HYDRAULIC CONDUCTIVITY K cm sec ft or m day ft or m yr Description Measure of the permeability of the transmissive layer To characterize concentrations in the transmissive layer representative measurements are required for the Darcy velocity or both the hydraulic flow gradient and the hydraulic conductivity of the flow system Representative measurements of the hydraulic conductivity of the transmissive layer should be obtained at one or more locations using appropriate slug test or pumping test methods Newell et al 2003 Typical Values Silts 1x10 6 4x1073 cm s Silty sands 1x105 4x1071 cm s Clean sands 1x10 3 1 cm s Gravels gt 1cm s Newell et al 1996 Source of Data Pump tests or slug tests at the site It is strongly recommended that actual site data be used for all matrix diffusion evaluations How to Enter Data 1 Select units and 2 Enter directly PARAMETER TRANSMISSIVE ZONE HYDRAULIC GRADIENT i Description The slope of the potentiometric surface In unconfined aquifers this is equivalent to the slope of the water table Typical Values 0 0001 0 1 ft ft 0 0001 0 1 m m MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 22 SRM DATA ENTRY Sour
75. applicable to LNAPL sites However we are not aware of any detailed research studies where matrix diffusion at LNAPL source zones was evaluated In addition some LNAPL components may persist for a long period of time making it difficult to understand whether the hydrocarbon plume is being sourced by matrix diffusion or from the persistent LNAPL phase Note that one group documented matrix diffusion effects associated with a MTBE TBA plume Rasa et al 2011 but this was not in an LNAPL source area Is the Toolkit able to simulate degradation in the low k zone Not at this time Numerical problems prevented a full implementation of the Dandy Sale Model with degradation Sale et al 2008b consequently this version of the Toolkit assumes no degradation in the low k zone However we hope to incorporate this feature in future versions of the Toolkit ESTCP MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y IV INTRODUCTION Over the past several years the groundwater research community in North America has become increasing aware that matrix diffusion has the potential to sustain dissolved contaminant concentrations in groundwater after the source is removed e g Chapman and Parker 2005 AFCEE 2007 Sale et al 2008a or after remediation removes or isolates contamination from transmissive compartments This persistent contaminant concentration can occur in the source zone itself or in some cases in the plume downgradient of the source
76. ater affinity of organic constituents for the organic carbon fraction of soil This value is chemical specific and can be found in chemical reference books Tetrachloroethene 155 mL g Benzene 66 mL g D Trichloroethene 93 mL g Ethylbenzene 204 mL g cis 1 2 Dichloroethene 29 mL g Toluene 140 mL g Vinyl Chloride 11 mL g Xylene 240 mL g 1 1 1 Trichloroethane 110 mL g MTBE 14 mL g TRRP 2008 Note that there is a wide range of reported values for example Mercer and Cohen 1990 report a Koc for benzene of 83 mL g For more information see Pankow and Cherry 1996 for solvents and Wiedemeier et al 1999 variety of constituents Source of Data Chemical reference literature such as Pankow and Cherry 1996 for solvents Wiedemeier et al 1999 variety of constituents or other references with chemical properties Alternatively one can use relationships between Koc and solubility or Ko and the octanol water partition coefficient Kow to determine Koc A collection of values is presented in the Chemical Parameter Database included in this manual How to Enter Data Enter directly Note that if the constituent is selected from the drop down list the Toolkit provides a value for the parameter MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 26 SRM DATA ENTRY Step 5 Plume Characteristics Key point about modeling area and concentration You do not need to model the entire source or plume area but only the ones exposed to the
77. ation OK to Override Next Highest Concentration Zone Blue Box in Picture See eier Save paa imate Length Length of Blue B 7 10E 01 ow Gra pene CET P Return to Model Selection Screen Return to Main Screen Approximate Width Width of Blue Box 3 00E 02 Figure 2 5 SRM Input Parameters Bromide MATRIX DIFFUSION TOOLKIT v USER S MANUAL Y 140 STUDY 2A SAND TANK STUDY SQUARE ROOT MODEL Time day 60 A Observed Bromide 2 Not Calibrated Lage 3 E c o D im Lol E L o E 6 Q Figure 2 6 Comparison of SRM Green Lines against Observed Concentrations Bromide The dark green line represents output using initial parameters The light green line represents the calibrated comparison Although visually not as good a match as the fluorescein data the overall trend in the data matched within an order of magnitude MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 141 CASE STUDY 2B SAND TANK STUDY DANDY SALE MODEL B Dandy Sale Model DSM Input Data Data Type Source of Data Hydrogeology Trans zone description sand Sand tank construction Trans zone effective 0 45 Sand tank construction porosity Low k zone description clay Sand tank construction Low k zone porosity 0 60 Estimated sand tank study Trans zone seepage 3 36E 4 cm sec Experimental sand tank study velocity Transport Key constituent fluorescein
78. ber times yields a probability distribution from which statistical characteristics such as mean percentile and variance can be obtained The Toolkit performs 1000 iterations for the Monte Carlo approach How to Enter Data 1 Specify a probability distribution for each parameter see Appendix A 3 of the User s Manual for details on probability distributions The Toolkit assumes that the values entered in the Input screen are the mean values For the normal distribution specify the standard deviation as a percent of the mean For lognormal distributions specify the error factor EF the ratio of the 95 percentile to the median of the lognormal data or the ratio of the median to the 5 percentile NOTE the error factor MUST be greater than one For uniform distribution specify the lower and upper limits as percentages of the mean 3 Perform Input Uncertainty Analysis MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 41 SRM ADVANCED UNCERTAINTY ANALYSIS Uncertainty Analysis Results PARAMETER SEE MASS DISCHARGE RESULTS Description The 5 percentile median and 95 percentile for mass discharge based on the user s choice of interpolation method and uncertainty in the input variables as defined by their probability distributions means variances and ranges Negative mass discharge values represent diffusion into the low k zone from the transmissive zone Positive values represent diffusion from the low k zone into the tran
79. cal Values 0 0001 10 000 mg L Source of Data Standard Method this is not needed Contour Map Method Use a contour map from the highest concentration period where groundwater samples were collected For example if concentrations have been decreasing use a concentration contour map from 1990 and not 2012 How to Enter Data Enter directly MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 31 SRM DATA ENTRY PARAMETER REPRESENTATIVE CONCENTRATION C31 Description Representative historical loading concentration of first modeled area denoted by the black box in the Toolkit input screen figure This value is a key parameter that can be changed during the calibration process to increase or decrease the simulated mass discharge concentration or mass to better match field data see the beginning of this section Typical Values 0 0001 20 000 mg L Source of Data Standard Method Data Source 1 Site History or Process Information For example the effective solubility of a constituent in a known DNAPL pool in the source could be used when modeling the source zone or if the DNAPL in the pool was comprised of 50 Trichloroethene TCE a concentration of 550 mg L 50 of TCE solubility of 1100 mg L could be used Alternatively one could use an estimate of the average historical concentration from the time the source started to the end of the loading period sometimes a groundwater model with a source decay term such as REMChlor
80. cal plane source produces a plume in the transmissive zone that loads up the low k zone due to diffusion This vertical plane source is shut off and diffusion results in a release of contaminants from the low k zone Assumptions The model makes the following assumptions 1 A source considered to be a thin pool is introduced at the contact between the two layers upgradient of x 0 2 A loading period occurs where there is a constant concentration of contaminants in the transmissive zone that drives contaminants into the low k zone 3 A release period occurs where the transmissive zone is assumed to have no concentration and an upper range estimate of release from the low k zone is generated 4 There is no degradation in either layer 5 Both layers are uniform homogeneous isotropic and infinite in the z direction perpendicular to groundwater flow 6 One dimensional 1 D advective transport in the transmissive layer parallel to the boundary of the layers is accompanied by transverse dispersion and diffusion 7 There is no longitudinal dispersion in the transmissive layer 8 1 D transverse diffusion transport occurs in the low k layer 9 Retardation of contaminants in both layers is based on instantaneous equilibrium between aqueous and sorbed phases Summary Active Source Sorbed phase mass in the low k layer at any time t can be calculated as pK MO Myu 2 m where M t Sorbed phase mass in the lo
81. can take some of the key conditions presence of a low k compartment contaminant solubility groundwater velocity sorption and time since the release occurred and make quantitative predictions about the concentration and or mass discharge that may remain in groundwater after all other source terms are removed In other words these MATRIX DIFFUSION TOOLKIT v USER S MANUAL Y 1 INTRODUCTION equations a simple mass discharge model and a more sophisticated analytical solution can be used to help answer these questions e What is the potential contaminant concentration in the source zone after the source material in the transmissive compartment is largely removed e What is the potential contaminant concentration downgradient in the plume after the source is removed or isolated such as with a slurry wall or Permeable Reactive Barrier To better equip the groundwater community with accessible useable and practical models for evaluating matrix diffusion effects the Environmental Security Technology Certification Program ESTCP of the U S Department of Defense DoD has funded the development of this Matrix Diffusion Toolkit Based on the Microsoft Excel platform the Toolkit is an easy to use comprehensive free software tool that can assist site personnel to effectively and efficiently estimate what effects matrix diffusion will have at their site and transfer the results to stakeholders Furthermore the software can assist proje
82. ce of Data Calculated by constructing potentiometric surface maps using static water level data from monitoring wells and estimating the slope of the potentiometric surface How to Enter Data Enter directly Step 4 Transport Low k Zone PARAMETER KEY CONSTITUENT Constituent of interest How to Enter Data Enter directly or choose from drop down list PARAMETER MOLECULAR DIFFUSION COEFFICIENT IN FREE WATER Do eH A factor of proportionality representing the amount of substance diffusing Description s EA Nar across a unit area through a unit concentration gradient in unit time Typical Values Benzene 9 8E 06 cm s Tetrachloroethene 8 2E 06 cm s Ethylbenzene 7 8E 06 cm s Trichloroethene 9 1E 06 cm s Toluene 8 6 06 cm s cis 1 2 Dichloroethene 1 1E 05 cm s Xylene 8 5E 06 cm s Vinyl Chloride 1 2E 05 cm s MTBE 9 4E 05 cm s 1 1 1 Trichloroethane 8 8E 06 cm s TRRP 2008 Note that there is a wide range of reported values for example Wiedemeier et al 1999 report a D for benzene of 1 1E 05 cm s For more information see Pankow and Cherry 1996 for solvents and Wiedemeier et al 1999 variety of constituents Source of Data Chemical reference literature such as Pankow and Cherry 1996 for solvents Wiedemeier et al 1999 variety of constituents or other references with chemical properties How to Enter Data 1 Select units and 2 Enter directly Note that if the constituent is selecte
83. ct managers in determining if remediation goals are achievable in the short term The Toolkit can be applied to virtually any site with heterogeneity in the subsurface DNAPL and or where persistent groundwater contaminant concentrations have been observed after source zone remediation The Toolkit provides a valuable tool for developing site conceptual models supporting site characterization efforts planning remedial designs and determining if matrix diffusion will affect remediation goals for groundwater sites The software can assist site personnel in updating or creating a more accurate conceptual site model which will enable them to determine if matrix diffusion processes are significant enough to cause rebounding of downgradient plume concentrations above remediation goals after plume remediation or isolation is complete Having this information available before a remedy is implemented could assist site stakeholders in selecting more appropriate remedies and effectively and efficiently addressing the potential issues of matrix diffusion with regulators Furthermore addressing extended remediation time frames caused by matrix diffusion would lead to savings in project costs The Toolkit provides the following tools to calculate and evaluate matrix diffusion effects 1 Square Root Model A module to provide planning level estimates of the mass discharge in units of grams per day caused by release from a low k diffusion dominated unit
84. d from the drop down list the Toolkit provides a value for the parameter PARAMETER APPARENT TORTUOSITY FACTOR EXPONENT p Description The Apparent Tortuosity Factor t relates the molecular diffusion coefficient in free water Do of a constituent in a porous medium to its effective diffusion coefficient De Values of t can range between 0 and 1 Estimations of t can be obtained using the relationship Where is the porosity and p the Apparent Tortuosity Factor Exponent Depending on the geologic medium values for p can vary between 0 3 and 5 4 Charbeneau 2000 Pankow and Cherry 1996 Dullien 1992 Lerman 1979 and Millington and Quirk 1961 MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 23 SRM DATA ENTRY Typical Values Clay Fractured Sandstone Granite Sandstone Shale Silt Payne et al 2008 Source of Data Literature How to Enter Data Enter directly Note that if the low k zone description is selected from the drop down list the Toolkit provides a value for the parameter PARAMETER RETARDATION FACTOR R Description The retardation factor is the ratio of the dissolved plus sorbed constituent mass to the dissolved constituent mass in the aqueous phase in a unit volume of aquifer The retardation factor is a function of both aquifer and constituent properties Typical Values For transmissive zones these retardation factors are commonly observed 1 3 typical for BTEX 2 5 typical for chlor
85. deposits A collection of default values is presented in the Geologic Parameter Database included in this manual How to Enter Data Enter directly Note that if the transmissive zone description is selected from the drop down list the Toolkit provides a default value for the parameter MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 46 DSM DATA ENTRY PARAMETER LOW k ZONE DESCRIPTION Description of the low k zone How to Enter Data Choose from drop down list or enter directly PARAMETER LOW k ZONE TOTAL POROSITY n Description Dimensionless ratio of the volume of voids to the bulk volume of the surface soil column matrix but excluding secondary porosity fractures solution cavities etc Total porosity is the ratio of all voids including non connected voids to the bulk volume of the aquifer matrix Effective porosity and any porosity data with secondary porosity information should not be used Typical Values The model input screen has these default values Clay 0 47 mid range of values below Silt 0 48 mid range of values below Sandstone shale 0 10 Pankow and Cherry 1996 Table 12 2 Fractured Sandstone 0 08 Pankow and Cherry 1996 Table 12 2 Granite 0 006 Pankow and Cherry 1996 Table 12 2 Values for total porosity from Domenico and Schwartz 1990 in part from Davis 1969 and Johnson and Morris 1962 SEDIMENTARY Porosity Gravel coarse 0 24 0 36 Gravel fine 0 25 0 38 Sand course 0 31 0
86. discussed in Appendix A 2 Guidelines for selecting key input parameters for the model are outlined in Dandy Sale Model Entry For help on results see Dandy Sale Model Results Uncertainty Analysis Uncertainty in mass flux estimates is a key issue in simulations of groundwater systems The Toolkit provides two options for analyzing this uncertainty One option performed automatically provides a lower range most likely value and an upper range for estimated outputs based on the specified source area concentrations The second option Advanced Uncertainty Analysis utilizes a Monte Carlo type approach to analyze uncertainty in the actual source concentration porosity apparent tortuosity factor exponent and retardation factor measurements With this tool groundwater practitioners can estimate the accuracy of the hydrologic measurements that are being used for the matrix diffusion calculation Monte Carlo analysis is a method of analyzing and quantifying uncertainties in model outputs due to the uncertainties in the input parameters Rong et al 1998 Monte Carlo analysis refers to a computer based system that uses random numbers from a probability distribution to obtain an approximation for the parameter of interest USEPA 1997 Bergin and Milford 2000 In the standard Monte Carlo approach simple random sampling and a large number of runs typically 100 to 1000 are required to obtain a meaningful probability distribution for the para
87. e removal time of 1996 was used initially in the Toolkit However the exact history of the source concentration is unknown therefore this parameter was used as a calibration parameter Specifically initially the source concentration was assumed to be constant at 475 mg L average of the estimated vs time curve from Chapman and Parker 2005 for 44 years and then turned off Figures 1 7 1 8 and 1 12 During the calibration process the solubility limit of TCE was used as the concentration and this time period was adjusted to better match the observed concentrations at location ML 10 in the year 2000 Figures 1 9 1 10 1 11 and 1 12 Monitoring data from the low k zone at locations WCP 70 and WCP 71 inside the sheet pile enclosure and location ML 10 were used for comparison to simulated source and plume concentrations respectively MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 122 CASE STUDY 1B INDUSTRIAL SITE CONNECTICUT DANDY SALE MODEL KEY POINTS The DSM of the Toolkit was able to reproduce observed groundwater concentrations within an order of magnitude Use of site specific values documented by Chapman and Parker 2005 and Toolkit default values for parameters with no site specific information provided a reasonable comparison to actual observed TCE concentrations in the source zone Figure 1 6 Therefore no adjustment of any input parameters was necessary A comparison using the observed reported maximum source co
88. e Method Find the highest concentration contour line on the historical plume map between the upgradient and downgradient transects denoted by the black box in the Toolkit input screen figure Estimate the area in square feet or square meters between these transects and inside this contour line At most sites you can get a close enough value by estimating the approximate width and approximate length of an equivalent area Enter these into the Toolkit Step 5 3 Standard Method Enter your own length and width for the second modeled area Lz and W2 Contour Line Method Find the second highest concentration contour line on the historical plume map between the upgradient and downgradient transects denoted by the blue box in the Toolkit input screen figure Estimate the area in square feet or square meters between these transects and inside this contour line At most sites you can get a close enough value by i estimating the approximate width and ii approximate length of an equivalent area The Toolkit will automatically subtract out overlapping areas If you don t want to use this second area set the length width and concentration equal to the values for the black box in the SRM input screen MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 28 SRM DATA ENTRY Determining loading concentration A loading concentration is required to run the model for up to two modeled areas the black box and the blue box on the input screen
89. e Source Aqueous phase mass in the low k layer at any time t can be calculated as Ma XF Cz OW 0 where MA Aqueous phase mass in the low k layer at time t M C x Z t 2 Aqueous concentration at lateral distance x depth z and time t M L calculated using Appendix A 2 1 Equation 1 i Cell of the concentration in the lateral distance from the source vs depth in low k output array N Total number of cells in the output array l Length of cell in the concentration in the lateral distance from the source vs depth in low k output array L MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 92 APPENDIX A 2 DANDY SALE MODEL h Height of cell in the concentration in the lateral distance from the source vs depth in low k output array L n Porosity of low k layer unitless and W Source zone width L Exhausted Source Once the source is exhausted the low k aqueous phase mass can be calculated at any time t as Matt XP C xz to lWhn 2 where M ag t z Aqueous phase mass at time t after the source has depleted M and C x zt x Aqueous concentration at lateral distance x depth z and time t after the source has depleted M L calculated using Appendix A 2 1 Equation 11 MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 93 Appendix A 2 5 Low k Sorbed Mass Purpose Determine the low k sorbed phase mass output in the Dandy Sale Model of the Matrix Diffusion Toolkit Given The verti
90. e high permeability compartment T Q Porosity of low k zone unitless C Mean plume concentration above the low k compartment during the loading period M L A lt Area of low k compartment beneath the transmissive zone plume L R Retardation factor for low k compartment unitless and D Effective aqueous phase diffusion coefficient in the low k compartment L2 T This can be estimated as D 2 9 D where p is the apparent tortuosity factor exponent unitless and D the molecular diffusion coefficient in free water L T MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 80 APPENDIX A 1 SQUARE ROOT MODEL Appendix A 1 2 Estimation of Concentration in Transmissive Zone Purpose Determine the transmissive zone concentration output of the Square Root Model of the Matrix Diffusion Toolkit Given The Toolkit provides an instantaneous mass discharge from the entire area A during the release period Note this mass discharge from the entire low k zone is assumed to be transported instantaneously to the downgradient edge of the modeled area there is no advection or travel time component of the Square Root Model But because diffusion from a low k zone is typically much slower than the travel time in the transmissive zone multiple decades vs months or years this approximation should not cause too much problem for most simulations f travel time is an important part of the simulation try using the DSM model Assum
91. e key input data for the Dandy Sale Model in the Toolkit The Dandy Sale Model is more sophisticated than the Square Root Model and requires additional input data However it is based on the same conceptual model of a two layer system and a loading period followed by a release period Can I calibrate the matrix diffusion models in the Toolkit Yes but with the caution that groundwater monitoring data may represent a combination of residual contaminants from the original source even if the source has been remediated and from matrix diffusion So a careful evaluation of the field data that you would calibrate against is necessary to make sure you aren t calibrating to the wrong values See Square Root Model Data Entry and Dandy Sale Model Data Entry for more information on how to calibrate the models MATRIX DIFFUSION TOOLKIT v USER S MANUAL Y 13 FREQUENTLY ASKED QUESTIONS How does the Toolkit handle uncertainty For the Square Root Model we suggest that the Toolkit results are within an order of magnitude a factor of 10 While this seems a large range the results will provide useful information in context of the wide range of concentrations and mass discharge found in source zones e g see the paper Contaminant Plume Classification System Based on Mass Discharge by Newell et a 2011 So the model obtains information about whether you think the loading concentration has been stable or decreasing over time and then applie
92. e models in the Toolkit are based on simplifying assumptions and one of the most important is the mathematical assumption that you have an infinitely thick low k zone in other words you can t input the thickness of your low k zone In practice that means the low k zone should be at least 1 meter thick for sites where matrix diffusion has been occurring for several decades Thinner low k zones such as thin lenses and stringers can be modeled but with more uncertainty in the final results If you are dealing with thin units less than 3 meters thick you should check to see If your particular combination of input data low k layer thickness retardation factor source loading start source removed and see result time are the key factors results in a problem by running the Dandy Sale model If this model showed a lot of contaminant mass has penetrated into the assumed infinitely thick low k zone farther than the actual thickness of the low k zone at your site then your simulation will likely deviate from MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 7 FREQUENTLY ASKED QUESTIONS reality at some point in time However if most of the mass is shallower than the thickness you observe in the field then the simulation should work reasonably well Overall sites with very thin clay stringers and or lenses may be difficult to simulate unless the timeframes are very short What does the Matrix Diffusion Toolkit do Low k zones can serve as indirect
93. e to achieve MCLs after the DNAPL is all gone Do I need special sampling data from the low k zones If you want to learn more about the potential impacts of matrix diffusion or want planning level modeling results then the Toolkit can be applied without sampling data from the low k zones The Toolkit can provide useful information about the general trends or style of matrix diffusion effects but absolute values of the simulated results may vary considerably from actual field observations The accuracy of the modeling results will be increased if there are data from the low k zones that can be used to calibrate the Toolkit models How accurate are the results The two models utilized in the Toolkit are very simplified representations of an extremely complicated process and field conditions Therefore even with sampling data from the low k zones we consider the potential results as an order of magnitude range accuracy But at many sites this level of accuracy will still provide very useful information for site managers MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y l QUICK START What input data will need Some of the input data are similar to what is used for existing solute transport models e g Darcy groundwater velocity size of the modeled area information on when the source started etc Other input data may appear new to many users for example you ll need to estimate the tortuosity of the low k materials where matrix
94. e un calibrated parameter set was used for this analysis Hydrogeological data was entered in Section 2 transport parameters in Section 3 source zone characteristics in Section 4 and desired output information in Section 5 Site specific values as documented by Chapman et al 2012 were available for all parameters except bulk densities organic carbon partitioning coefficient and Transverse Vertical Hydrodynamic Dispersivity For these Toolkit default literature values were used Values of zero were used to calculate retardation factors of one To account for the travel of contaminated groundwater present at the time of the source removal an effective source removal time of 24 days was used in the Toolkit Transmissive zone concentrations output from the Toolkit were multiplied by eight to account for the eight interfaces of the four clay layers and 2 96 adjusted for the height of the tank compared to the Toolkit built in 10 ft well screen Comparison of the observed and simulated concentrations is provided below Simulation Time days 30 62 89 124 Observed Concentration mg L 0 076 0 0044 0 0018 0 0012 Simulated Concentration mg L 0 031 0 0036 0 0018 0 0010 Simulated Observed 0 40 0 82 0 97 0 85 KEY POINT The purpose of this evaluation was to see if the Toolkit DSM could simulate a difficult problem four very thin layers in a system with advection As described in the Uses and Limitation
95. e water Trans zone apparent 0 33 Literature Toolkit default tortuosity factor exponent Low k zone apparent 0 33 Literature Toolkit default tortuosity factor exponent Trans zone bulk density 1 7 g mL Site estimate Low k zone bulk density 1 5 g mL Site estimate Trans zone foc 0 03896 Site evaluation Low k zone foc 0 054 Site evaluation Organic carbon partitioning 93 5 L kg e Literature Toolkit default coefficient Coefficient of transverse hydrodynamic dispersion 0 001 m Literature Toolkit default Source Zone Source zone length 32 1 m Site map Characteristics Source zone width 39 3 m Site map Source loading starts in year 1952 Site history Source removed in year Source zone evaluation 1997 Estimated from site history plume zone evaluation 1996 initial 1978 calibrated General See results for year 1997 source zone evaluation Site monitoring data 2000 plume zone evaluation Lateral distance from source 0 001 m source zone Site map evaluation 280 m plume zone evaluation Vertical depth 3 m MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 121 CASE STUDY 1B INDUSTRIAL SITE CONNECTICUT DANDY SALE MODEL DSM Summary The Toolkit DSM was used to estimate TCE groundwater concentrations in the low k zone following DNAPL remediation at an industrial facility The DSM was used to estima
96. ear 1997 Concentration in Low k Zone mg L 0 00E 00 2 00E 02 4 00E 02 6 00E 02 8 00E 02 1 00E 03 0 0 1 1 E o o NI X z o er e 2 IE m 2 o a Log C NE Restore Original Graph See Results for Lateral Distance from Source x 0 004 m Update Graph l Low Permeability Zone See 2 D Low k Aqueous Conc See Trans Zone Aqueous Conc a Export Low k 2 D Data HELP See Low k Aq Conc vs Dist See Trans Zone Mass Discharge Sorbed Mass CALEX kg Return to DSM Data Input Save Data See Trans Zone Sorbed Conc Total Mass PAN e yd kg See 2 D Low k Sorbed Conc See Trans Zone Total Conc Return to Main Screen See 2 D Low Perm Total Conc Figure 1 5 DSM Output Source Area Low k Zone Concentrations MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 125 CASE STUDY 1B INDUSTRIAL SITE CONNECTICUT DANDY SALE MODEL Using Cs 1300 mg L E w E o N e o gel E z Kai Qa o a 600 900 1200 1500 TCE Concentration mg L Figure 1 6 Comparison of DSM Source Area Low k Concentrations Green and Purple Lines against Observed Concentrations in WCP 70 and WCP 71 in 1997 Based on Figure 6a of Chapman and Parker 2005 The match between actual and modeled results is very close MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 126 CASE STUDY 1B INDUSTRIAL SITE CONNECTICUT DANDY SALE MODEL DSM Data Input Screen DATA_INPUT INSTRUCTIONS LIC Enter va
97. ed on instantaneous equilibrium between aqueous and sorbed phases Summary Active Source Since the medium is saturated with water the water content equals the porosity Consequently the total concentration mass of the constituent per unit bulk volume can be obtained using U 1 n phK Caen d CG z 0 3059 1 b where C rotai X Z t Total concentration at lateral distance x depth z and time t MIMI C x z t Aqueous concentration at lateral distance x depth z and time t M L calculated using Appendix A 2 1 Equation 1 n Porosity of low k layer unitless p Bulk density of low k layer M L Ka Soil water partitioning coefficient L M MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 90 APPENDIX A 2 DANDY SALE MODEL Tac Koc foc Fraction organic carbon of the low k layer unitless and Koc Organic carbon partitioning coefficient L M Exhausted Source Once the source is exhausted the low k total concentration can be calculated at any time tas Cri te7 t T C x zZ t t ntepa 2 where C ota X Z t t Total concentration at lateral distance x depth z and time t after the source has depleted M M and C x z t x Aqueous concentration at lateral distance x depth z and time t after the source has depleted M L calculated using Appendix A 2 1 Equation 11 MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 91 APPENDIX A 2 DANDY SALE MODEL Appendix A 2
98. el output can be compared to actual field data from monitoring wells using either a concentrations comparison or a mass discharge comparison Most times the initial run will not produce modeled data that match field data Considerations and recommended steps to improve the fit of simulated data to field data are provided below The first caveat associated with calibrating the DSM is that the model assumes the original source zone is completely cleaned up and does not account for any residual source In other words at many sites the concentrations from matrix diffusion may only be causing part of the contaminant concentrations in monitoring wells Consequently an exact match to observed concentration in a monitoring well should not be attempted if there is any uncertainty in matrix diffusion processes being the sole source of contaminants in the modeled zone In cases where a good comparison between concentrations and or mass discharge from actual groundwater monitoring data can be made either because there is no residual source or the matrix diffusion signal can be abstracted out the recommended sequence of model input values to change is MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 44 DSM DATA ENTRY 1 First change the Plume Loading Concentration Co If the simulated concentrations are higher than observed concentrations reduce the representative concentrations 2 If it is still difficult to get a good fit try changing either
99. elocity in the transmissive layer and t the time of interest 3 Otherwise divide the spatial distance x into N subdivisions 4 Calculate x N x At VN 6 Determine the initial mass loaded onto the reference volume over the time period At Co Mremain VWndt where W is the source zone width 7 Loop estimated losses from the reference volume over the N spatial subdivision That is for 1 to N repeat a Determine the midpoint for each spatial subdivision Ax xx i DAx b Determine the midpoint of the mass lost to the low k zone at each spatial subdivision At tt t N DAt gt c Using Equation 2 calculate the flux across the interface at spatial location xx and time tt d Remove mass from the reference volume at each spatial subdivision over the period At Mremain Mremain Jy xx tt AxWAt e Calculate the concentration in the hypothetical well by dividing the mass remaining by the volume of water in the reference volume Mremain Cwell RWAxh n where hy is the screen interval of the well 8 The concentration in the transmissive layer at distance x and time t Cwen at i N Exhausted Source MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 102 APPENDIX A 2 DANDY SALE MODEL Concentration in the transmissive zone after the source is exhausted can be determined using the principle of superposition For this purpose 1 Well concentration is calculated for time t usin
100. eous and sorbed phases Summary In 2008b Sale et al developed Equation 1 to calculate the concentration in the transmissive zone with an active source b2x x b b Crans 5 Z t co Zeg 2 vx 2 e e bzerf E VX 23 bs emd Z ZP 25 QZ ba xe 2 sf JE sch o rol 5 Ca ZS However as demonstrated by Bolhari 2012 the above equation has a finite domain of application lt 100 meters Consequently the Toolkit estimates the transmissive zone aqueous concentrations by determining the flux across the low k transmissive zone interface and assuming discharge to a hypothetical well with a 10 ft 3 m screened interval This 10 foot screened interval was selected because at an actual field site contamination diffusing from a low k zone might spread vertically above a 1 foot screen It was thought to be very unlikely that there would be more than 10 feet of vertical spreading in the transmissive zone Bottom line the 10 foot screened interval is hard wired into the model and cannot be changed Due to computational limitations all transmissive zone solutions show increasing numerical imprecision for lengths greater than 1500 m Therefore we recommend limiting the lateral distance from the source to S 1500 m for any model runs involving transmissive zone solutions If you are sure all the mass discharge is being captured by a well with a different screened interval you can get the modeled concentration in this well by multi
101. er day Seyedabbasi et al 2012 This helps support the contention that there are a number of Late Stage chlorinated solvent sites where DNAPL is a relatively small part of the source and matrix diffusion is the predominate contributor Sale ef al 2008a b Obviously if there is a very large DNAPL release of hundreds of thousands of pounds then DNAPL will likely be a large part of the site conceptual model for a long period of time e For the remediation case there are perhaps thousands of sites where active in situ remediation has removed DNAPL from the transmissive zone but has left behind contaminants in the low k zones These sites are likely to be dominated by matrix diffusion effects now or sometime in the near future What is a low k zone Do I have these zones at my site Based on her research program at the University of Guelph Dr Beth Parker has a rule of thumb indicating that matrix diffusion can be an important process if there is a plume in a transmissive zone that is in contact with adjacent zones that have permeabilities lower than by a factor of 100 or more In other words if a contaminant plume moving in a 10 cm sec sand is in contact with a 10 cm sec silt then the silt can be charged up with contaminants during a loading period when concentrations in the sand are higher than the silt and then slowly discharge contaminants into the sand via diffusion when the silt has higher concentrations than the sand Th
102. es for example Mercer and Cohen 1990 report a Koc for benzene of 83 mL g For more information see Pankow and Cherry 1996 for solvents and Wiedemeier et al 1999 variety of constituents Typical Values Source of Data Chemical reference literature such as Pankow and Cherry 1996 for solvents Wiedemeier et al 1999 variety of constituents or other references with chemical properties Alternatively one can use relationships between Koc and solubility or Ko and the octanol water partition coefficient Kow to determine Koc A collection of values is presented in the Chemical Parameter Database included in this manual How to Enter Data Enter directly Note that if the constituent is selected from the drop down list the Toolkit provides a value for the parameter Step 4 Source Zone Characteristics PARAMETER SOURCE ZONE LENGTH L Description Estimated length of the original source zone parallel to groundwater flow that is upgradient of the modeled area This length is only used to establish a parameter that exponentially reduces the vertical concentration in the vertical source plane by the W in the figure below Close to the bottom of the vertical plane source the concentration is equal to Co page 42 at the top of the vertical plane the concentration is lower based on equation 3 on page 78 which uses o page 48 which in turn is a function of L You can ignore L and just enter your own value of o if you p
103. esssr veces cce ese ee eeeeee enorer errer renen 42 DANDY SALE MODEL DATA ENTRY 00 cccceceeseeeeee cece eee reenn ennenen nnn nnes 44 DSM Data Input Screen d ENNERT ENEE SEENEN IRR RR RE 46 Step 1 System UNIS nui ai ei Er aer LET a Le eR LR EN n e ANES EAE itt 46 otep 2 Hydrogeology E WE HR RE BEP E EE APR 46 Slepi3 Nee EE 49 Step 4 Source Zone Characteristics eee ee eee 53 SLED ON EE 56 DSM Models ROSUIES x20 sate e ter ooa ed aae ER ee ee aea 58 NAPL DISSOLUTION MODEL DATA ENTRY cccceseeeceeseseeeeeeesneeeeseeeeesesesneeseseeeeeesesesneesenesnanen 61 NAPL Dissolution Model Data Input Screen 62 Step SYStem UNIS ites toto o Rea Laeti 62 Step 2 Hydrogeology Transmissive Zone 62 Step 3 Transport ope b ne reed EL ea eda e eR Ee PUR A Read tog uto etude 64 MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y ii TABLE OF CONTENTS Step 4 Plume Characteristics sse eee eee eee eee DD NAPL Dissolution Model Re SU S T r e a ae e a E a eat a e 68 CHEMICAL PARAMETER DATABASE cscce eee essere eenn eenn 69 GEOLOGIC PARAMETER DATABASE ee nne ennenen enen 71 REFERENCES ege ede EENEG 73 MATRIX DIFFUSION TOOLKIT TROUBLESHOOTING TIPS sese eee sese errer eee 76 Minimum System Requirements esses enne renne ennenen e 76 Installation and Start Up 76 Spreadsheet Related Problems sss seien nennen nnns 76 Common Error Messages E 76 AGKNOWLEDGEMENTS irruere iter ence edere ENEE Ee
104. estimate mass How to Enter Data Enter directly PARAMETER NEXT STEP SHOW GRAPH Proceeds to the results of matrix diffusion analysis PARAMETER si ETER NEW NEW SITE CLEAR DATA DATA E Clears ALL data related to the SRM model in the Toolkit memory banks Use this button to start a new project PARAMETER PASTE EXAMPLE Description Clears ALL data related to the SRM model in the Toolkit memory banks and pastes an example dataset The example dataset used in the Toolkit is obtained from Chapman and Parker 2005 PARAMETER si ETER SAVEDATA e DATA PARAMETER _ Saves all the SRM model data Sa ATE e NOT ADD ANY EXTENSIONS TO FILE NAME WHEN SAVING PARAMETER ETER LOADDATA e DATA Description Loads data files saved through the Toolkit DO NOT EDIT ANY TOOLKIT FILES Editing files may cause the Toolkit to crash PARAMETER RETURN TO MODEL SELECTION SCREEN Returns to the Model Selection Screen MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 37 SRM DATA ENTRY PARAMETER RETURN TO MAIN SCREEN Returns to the Matrix Diffusion Toolkit Main Screen MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 38 SRM RESULTS SRM Model Results PARAMETER SEE MASS DISCHARGE RESULTS Description Mass discharge from the entire low k and transmissive zones in units of g day Negative mass discharge values represent diffusion into the low k zone from the transmissive zone Positive values represent release from the low k zone into t
105. ffectively you must enable the macros NOTE Although the Toolkit uses Microsoft Excel some information in the Apply Related Tools module calls Adobe Acrobat pdf documents Some features in the module may not work unless you have this program installed on your computer Spreadsheet Related Problems Backspace doesn t clear cell Use the delete key on the keyboard or the mouse to clear data The buttons won t work The Toolkit is built in the Excel spreadsheet environment and to enter data one must click anywhere outside the cell where data was just entered If you can see the numbers you just entered in the data entry part of Excel above the spreadsheet the data have not yet been entered Click on another cell to enter the data is displayed in a number box The cell format is not compatible with the value e g the number is too big to fit into the window To fix this select the cell pull down the format menu select Format Cells and click on the Number tab Change the format of the cell until the value is visible If the values still cannot be read select the format menu select Cells and click on the Font tab Reduce the font size until the value can be read DIV 0 is displayed in a number box The most common cause of this problem is that some input data are missing In some cases entering a zero in a box will cause this problem Double check to make certain that data required for your run have been entered in all
106. fusion Toolkit developed for the Environmental Security Technology Certification Program ESTCP by GSI Environmental Inc Houston Texas CONTACTS Dr Shahla Farhat GSI Environmental Inc 713 522 6300 skfarhat gsi net com Dr Charles Newell GSI Environmental Inc 713 522 6300 cjnewell gsi net com MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y i TABLE OF CONTENTS MATRIX DIFFUSION TOOLKIT Environmental Security Technology Certification Program CONTENTS Section Page No QUICK STAR RE l INTRODUG TION fec 1 INTENDED USES FOR MATRIX DIFFUSION TOOLKIT AND LIMITATIONS 4 FREQUENTLY ASKED QUESTIONS a enen 7 MATRIX DIFFUSION TOOLKIT MODELS eeeereeeeneeenee nnne nnne nnne nnn nennen 15 Square Root Model SRM sss esee enirn nennen enne nnne nnns 15 Dandy Sale Model OM 15 Uncertainty Analysis du oe Lo Hexe aa iy dava Re RE e RUD a ea Eee ieee AERE a 16 SQUARE ROOT MODEL DATA ENTRY esse sees essere ennenen ennenen eenn ennenen 18 SRM Data Input SCre TTT 20 SLOP TE System Units iae de une rl n de Ra ce A E ce 20 ee 20 Slop 3 HYOGO e 21 Step 4 Transport Low k Zone 23 Step 5 Plume Characteristics eee eee eee 27 TT eT 34 Step 7 Field Data for Comparison streo a eaaa a aaa eaaa aaa aa aO a aa 35 SRM lee ERT 39 Uncertainty Analysis Perform Uncertainty Anaveis e eee eee sees 41 Uncertainty Analysis Results ase
107. g steps 1 through 8 above 2 Well concentration is calculated for time t gt r where zis the source persistence time i e the time in which the source is active as a Assume a monitoring well with a 10 ft 3 m screened interval located at the distance x of interest b Divide the spatial distance x into N subdivisions c Calculate N x At V N e Determine the initial mass loaded onto the reference volume over the time period At Co Mremain p YWrAt where W is the source zone width f Loop estimated losses from the reference volume over the N spatial subdivision That is for 1 to N repeat a Determine the midpoint for each spatial subdivision Ax xx i 1 Ax b Determine the midpoint of the mass lost to the low k zone at each spatial subdivision i For x 2 V t 7 EE i V ii For x lt V t 7 xx x err c Using Equation 2 calculate the flux across the interface at spatial location xx and time tt d Remove mass from the reference volume at each spatial subdivision over the period At Mremain Mremain Jy xx tt AxW At MATRIX DIFFUSION TOOLKIT v USER S MANUAL Y 103 APPENDIX A 2 DANDY SALE MODEL e Calculate the concentration in the hypothetical well by dividing the mass remaining by the volume of water in the reference volume Cwell RW Axh n where hw is the screen interval of the well g The concentration in the transmissive layer at distance
108. geneous isotropic and infinite in the z direction perpendicular to groundwater flow 6 One dimensional 1 D advective transport in the transmissive layer parallel to the boundary of the layers is accompanied by transverse dispersion and diffusion 7 There is no longitudinal dispersion in the transmissive layer 8 1 D transverse diffusion transport occurs in the low k layer 9 Retardation of contaminants in both layers is based on instantaneous equilibrium between aqueous and sorbed phases Summary Aqueous phase mass in the low k layer at any time t can be calculated as Maq t KE CweulW hyn where M t Aqueous phase mass in the transmissive layer at any time t M Cwen Well concentration at lateral distance x and any time t M L calculated using Appendix A 2 7 N Total number of wells in the output graph e intervals on x axis of graph I Distance to midpoint on each side of well L hw Screen interval of well L and n Porosity of transmissive layer unitless MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 108 APPENDIX A 2 DANDY SALE MODEL Appendix A 2 11 Transmissive Layer Sorbed Mass Purpose Determine the transmissive layer sorbed phase mass output in Dandy Sale Model of the Matrix Diffusion Toolkit Given The vertical plane source produces a plume in the transmissive zone that loads up the low k zone due to diffusion This vertical plane source is shut off and diffusion results in
109. he transmissive zone Note this mass discharge from the entire low k zone is assumed to be transported instantaneously to the downgradient edge of the modeled area there is no advection or travel time component of the Square Root Model However because diffusion from a low k zone is typically much slower than the travel time in the transmissive zone multiple decades vs months or years this approximation should not cause too much problem for most simulations If travel time is an important part of the simulation try using the DSM model A lower range most likely value and an upper range for estimated outputs are provided based on the input source area concentrations The user may use the Log lt gt Linear button to see the results on a semi log plot PARAMETER WHAT S UP WITH THE GAP Description In this simple model the mass discharge due to release from low k zones in the first few seconds hours and days after the loading period ends is extremely high but only lasts a short time Consequently to avoid confusion associated with these high mass discharge spikes a 1 yr gap between the loading period termination and the start of the release period is utilized in the output graph Matrix diffusion is a long process typically decades or more Because the transition phase between the loading period and release period is a year or more at many sites such as the case where remediation is performed the missing year is not like
110. highest historical concentrations You can likely get 90 of the loading from matrix diffusion by modeling the area inside the two highest concentration contour lines as shown by the blue and black boxes in the figures below on a historical plume map in the area of interest Two methods are provided for this Step First time users are more likely to use the Contour Map Method but skilled users will likely go straight to the Standard Method Standard Method Determine the area you want to model for matrix diffusion and enter the length width and representative historical loading concentration Note the model allows you to enter data for two different areas at your site i e two lengths two widths and two representative loading concentrations Contour Map Method Use a method based on lengths widths and concentrations from a historical contour map preferably one with the highest historical concentrations observed during the monitoring record For example if source concentrations have been decreasing over time use a concentration contour map from 1990 and not 2012 Determining modeling length and width The first goal is to define a length width and loading concentration for the first modeled area black box and the second modeled area blue box excluding the black box area Here are two options for entering the data Step 5 1 Standard Method Enter your own length and width in the model Contour Map Method Draw a downgradie
111. icult We recommend two ways to estimate the historical loading concentration 1 Historical Process Information At some sites you might have certain process knowledge about the modeling area during the loading period such as this area had DNAPL or there was a release of a certain strength waste In this case estimate the historical groundwater concentrations based on this information such as the effectively solubility of the contaminant in a DNAPL and use this as the Loading Concentration 2 Highest Observed Concentration More commonly you will not have process knowledge and in that case we recommend you use the highest observed concentration from a groundwater monitoring point in the modeled area the two boxes as a starting MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y I QUICK START point While not perfect this method is based on real data and represents observed loading concentrations in the modeled area Many sites have more than one low k layer in contact with the plume You can simulate this heterogeneity outside of the Toolkit using the default two layer one interface configuration and then multiplying the mass discharge from the Toolkit by the number of interfaces the number of contacts between a transmissive zone and low k zone see Inset 1 on page 10 You can do the same for concentration output if each interface intersects the screen of the assumed monitoring well Can the Toolkit be used with fractured rock sites
112. imensional 1 D advective transport in the transmissive layer parallel to the boundary of the layers is accompanied by transverse dispersion and diffusion 7 There is no longitudinal dispersion in the transmissive layer 8 1 D transverse diffusion transport occurs in the low k layer 9 Retardation of contaminants in both layers is based on instantaneous equilibrium between aqueous and sorbed phases Summary Using a linear soil water partitioning coefficient the sorbed concentration in the low k layer at any time f can be calculated as Csorbed x t Cweu Ka 1 where Csorbea X t Sorbed concentration at lateral distance x and any time t MIMI C well Well concentration at lateral distance x and any time t MI Calculated using Appendix A 2 7 Ka Soil water partitioning coefficient L M foc Koc foc Fraction organic carbon of the transmissive layer unitless and Koc Organic carbon partitioning coefficient L M MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 105 APPENDIX A 2 DANDY SALE MODEL Appendix A 2 9 Transmissive Layer Total Concentration Purpose Determine the transmissive layer total concentration output in the Dandy Sale Model of the Matrix Diffusion Toolkit Given The vertical plane source produces a plume in the transmissive zone that loads up the low k zone due to diffusion This vertical plane source is shut off and diffusion results in a release of contaminants from the low k zone A
113. inated solvents It is thought that retardation factors for low k zones are higher than transmissive zones Currently there are few sites where these values have been determined however Source of Data Usually estimated from soil and chemical data using the following expression R 1 Ka pa m where Kg Koc foc and pa bulk density n porosity Koc organic carbon water partition coefficient Kg distribution coefficient and foc fraction organic carbon on uncontaminated soil In some cases the retardation factor can be estimated by comparing the length of a plume affected by adsorption such as the benzene plume with the length of a plume that is not affected by adsorption such as chloride Most plumes do not have both types of constituents so it is more common to use the estimation technique See fraction organic carbon below for more information How to Enter Data 1 Select units and enter directly or 2 Calculate by pressing the Calculate R button and entering values for a Soil Bulk Density and b Distribution Coefficient or Fraction Organic Carbon and Organic Carbon Partitioning Coefficient PARAMETER SOIL BULK DENSITY OF LOW k ZONE rhob Description Density of the saturated low k zone referred to as soil excluding soil moisture MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 24 SRM DATA ENTRY Typical Values Although this value can be measured in the lab estimated values are used in m
114. ing period occurs where there is a constant concentration of contaminants in the transmissive zone that drives contaminants into the low k zone 3 A release period occurs where the transmissive zone is assumed to have no concentration and an upper range estimate of release from the low k zone is generated 4 There is no degradation in either layer 5 Both layers are uniform homogeneous isotropic and infinite in the z direction perpendicular to groundwater flow 6 One dimensional 1 D advective transport in the transmissive layer parallel to the boundary of the layers is accompanied by transverse dispersion and diffusion 7 There is no longitudinal dispersion in the transmissive layer 8 1 D transverse diffusion transport occurs in the low k layer 9 Retardation of contaminants in both layers is based on instantaneous equilibrium between aqueous and sorbed phases Summary Active Source Total mass in the low k layer at any time t can be calculated as Mig t Mag t M OU where M t Total mass in the low k layer at time t M Mall Aqueous phase mass in the low k layer at time t M calculated using Appendix A 2 4 Equation 1 and M t Sorbed phase mass in the low k layer at time t M calculated using Appendix A 2 5 Equation 1 Exhausted Source Once the source is exhausted the low k total mass can be calculated at any time t as MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 96 APPENDIX A 2 DANDY SALE
115. ing potentiometric surface maps using static water level data from monitoring wells and estimating the slope of the potentiometric surface MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 48 DSM DATA ENTRY Step 3 Transport PARAMETER KEY CONSTITUENT Constituent of interest How to Enter Data Enter directly or choose from drop down list PARAMETER PLUME LOADING CONCENTRATION IMMEDIATELY ABOVE LOW k ZONE IN VERTICAL PLANE SOURCE DURING LOADING PERIOD C lt Description Concentration used at base of vertical plane source see figure below from Sale et al 2008b Semi infinitte transmissive zone 6 9 sand N Semiinfinite low permeability zone e g j silt Source Co at the contact decaying exponentially with increasing distance from the interface Typical Values 0 0001 20 000 mg L Source of Data Data Source 1 Site History or Process Information For example the effective solubility of a constituent in a known DNAPL pool in the source could be used when modeling the source zone or if the DNAPL in the pool was comprised of 50 Trichloroethene TCE a concentration of 550 mg L 50 of TCE solubility of 1100 mg L could be used Alternatively one could use an estimate of the average historical concentration from the time the source started to the end of the loading period sometimes a groundwater model with a source decay term such as REMChlor Falta et al 2007 can be used to estimate hi
116. ion performed automatically provides a lower range mostly likely value and an upper range for estimated outputs based on the specified Source area concentrations The second option Advanced Uncertainty Analysis utilizes a Monte Carlo type approach to analyze uncertainty in the actual concentration porosity apparent tortuosity factor exponent and retardation factor measurements With this tool groundwater practitioners can estimate the accuracy of the hydrologic measurements that are being used for the matrix diffusion calculation PARAMETER SAVE DATA Description Saves all the SRM model data DO NOT ADD ANY EXTENSIONS TO FILE NAME WHEN SAVING Note that this option does not save any edits performed on the graphs by the user To save such edits use the save function of Excel and save the entire Toolkit file PARAMETER RETURN TO SRM DATA INPUT Returns to the SRM data input screen PARAMETER EXPORT PRINT DATA TABLE Description Exports the time mass discharge mass concentration and plume magnitude information shown in the table into a text file for use in other programs Prints the data table shown on the screen on the default printer To print on a different printer select the printer in the Print options in Excel and then press the Print button PARAMETER RETURN TO MAIN SCREEN Returns to the Matrix Diffusion Toolkit Main Screen MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 40 SRM ADVANCED UNCERTAINTY ANALYSI
117. ion at x 0 z 0 and H the Heaviside step function such that Oift lt t lift gt T H t 1 Numerical Integration Method The Toolkit employs a 10 pt Gaussian quadrature to solve polynomials MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 87 APPENDIX A 2 DANDY SALE MODEL Appendix A 2 2 Low k Sorbed Concentration Purpose Determine the low k sorbed concentration output in the Dandy Sale Model of the Matrix Diffusion Toolkit Given The vertical plane source produces a plume in the transmissive zone that loads up the low k zone due to diffusion This vertical plane source is shut off and diffusion results in a release of contaminants from the low k zone Assumptions The model makes the following assumptions 1 A source considered to be a thin pool is introduced at the contact between the two layers upgradient of x 0 2 A loading period occurs where there is a constant concentration of contaminants in the transmissive zone that drives contaminants into the low k zone 3 A release period occurs where the transmissive zone is assumed to have no concentration and an upper range estimate of release from the low k zone is generated 4 There is no degradation in either layer 5 Both layers are uniform homogeneous isotropic and infinite in the z direction perpendicular to groundwater flow 6 One dimensional 1 D advective transport in the transmissive layer parallel to the boundary of the layers is accompanied by
118. ioning Coefficient CASE STUDY 2B SAND TANK STUDY DANDY SALE MODEL DAIA INPUT INSTRUCTIONS Version 1 0 LLL Enter value directly Value calculated by Toolkit Do not enter data Transmissive Zone Low k Zone Results Calculate V L Calculated Here SOURCE ZONE CHARACTERISTICS Source Zone Length Source Zone Width Transverse Vertical Hydrodynamic DispersMity Source Loading Starts in Year Source Removed in Year GENERAL See Release Period Results for Year format mm Lateral Distance from Source x m Depth into Low k Zone zl G t Calculated R Figure 2 7 DSM Input Parameters Fluorescein shown for output time 30 days MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y E 144 CASE STUDY 3 FORMER DRY CLEANER FLORIDA Overview The Toolkit was used to estimate the effects of diffusion into and from low k zones at the former Building 106 area in Operable Unit 3 OU3 a former dry cleaner site at Naval Air Station NAS Jacksonville Florida studied by GSI Environmental and the University of Guelph The site was studied using University of Guelph high resolution core sampling techniques Mr Mike Singletary of the Naval Facilities Engineering Command was the Navy point of contact for this project The DSM was used for this analysis and applied as follows e Step 1 Due to a lack of historical information on the site Toolkit default values were used as initial parameters e Step 2 Toolkit outputs were
119. iption Exports the time mass discharge concentration and mass shown on the graphs into a text file for use in other programs PARAMETER RETURN TO MAIN SCREEN Returns to the Matrix Diffusion Toolkit Main Screen MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 43 DSM DATA ENTRY DANDY SALE MODEL DATA ENTRY DSM Data Input Screen DSM Results Three important considerations regarding data input are 1 To see the example dataset in the input screen of the software click on the Paste Example button on the lower right portion of the input screen The example dataset used in the Toolkit is obtained from Chapman and Parker 2005 2 Because the Toolkit is based on an Excel spreadsheet you have to click outside of the cell where you just entered data or hit Return before any of the buttons will function Additionally REMOVING OR ADDING rows or columns in input screens may cause the program to crash 3 Parameters used in the model are to be entered directly into the white blue cells NOTE Although literature values are provided site specific hydrogeologic transport and plume characteristic values will likely provide better results If literature values are used and there is uncertainty in the value chosen sensitivity analyses should be conducted to determine the effects of the uncertainty on model predictions Recommendations Regarding Calibrating Fitting the DSM to Actual Field Data After the model has been set up and run mod
120. ix diffusion or want planning level modeling results the Toolkit can be run without data from the low k zones MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 9 FREQUENTLY ASKED QUESTIONS INSET 1 HOW DO I GO FROM THIS Depth meters T l TO THIS Basic Idea Step 1 Count the number of interfaces where a silt or clay layer that is at least 1 meter thick is in contact with the plume in transmissive sands gravels For the example above if there was a plume in the yellow Sand Unit 1 between 10 and 12 meters there would be two interfaces both the clays above and below Sand Unit 1 are at least 1 meter thick the model cannot simulate very thin clay stringers or lenses See the FAQs Step 2 Run either model in the Toolkit and get the result you are interested in mass discharge grams per day mass kilograms or concentration in a well mg L Step 3 Multiply the results in Step 2 by the number of interfaces from Step 1 For example if the Toolkit determines your simulation has a mass discharge of 2 grams per day gpd and you have 2 interfaces the end result is a mass discharge of 4 gpd for your site Similarly if the Toolkit determines the concentration is 0 51 mg L and you have 3 interfaces the actual concentration is 1 53 mg L See the four examples below for further details ser TEL Plume in transmissive zone in contact with 1 interface Low k unit is gt 1 meter thick
121. kit provides a value for the parameter MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 50 DSM DATA ENTRY PARAMETER LOW k ZONE APPARENT TORTUOSITY FACTOR EXPONENT p Description The Apparent Tortuosity Factor t relates the molecular diffusion coefficient in free water Do or Dag of a constituent in a porous medium to its effective diffusion coefficient De Values of t can range between 0 and 1 Estimations of t can be obtained using the relationship T P Where is the porosity and p the Apparent Tortuosity Factor Exponent Depending on the geologic medium values for p can vary between 0 3 and 5 4 Charbeneau 2000 Pankow and Cherry 1996 Dullien 1992 Lerman 1979 and Millington and Quirk 1961 Typical Values Clay Fractured Sandstone Granite Sandstone Shale Silt Payne et al 2008 How to Enter Data Enter directly Note that if the low k zone description is selected from the drop down list the Toolkit provides a value for the parameter PARAMETER BULK DENSITY OF TRANSMISSIVE ZONE pp Description Density of the saturated transmissive zone aquifer material referred to as soil excluding soil moisture Typical Values Although this value can be measured in the lab in most cases estimated values are used A value of 1 7 g mL is used frequently Source of Data Either from an analysis of soil samples at a geotechnical lab or more commonly application of estimated values such as 1 7 g mL How
122. le Florida SYSTEM UNITS issi z DNAPL R Transmissive Zone SI Units English Units Source Deis HYDROGEOLOGY l Transmissive Zone Description Sand Sand Transmissive Zone Effective Porosity 0 25 Low k Zone Low k Zone Description Clay Clay X Low k Zone Total Porosity 0 47 Transmissive Zone Seepage Velocity V 2 50E 01 ft vr KA Calculate V 2 een Here TRANSPORT Key Constituent enter directly or choose from drop down list PCE PCE x Plume Loading Concentration Immediately Above Low k Zone in Vertical Plane Source During Loading Period 143 moh EN Molecular Diffusion Coefficient in Free Water 4 20E 10 m2 sec z Transmissive Zone Apparent Tortuosity Factor Exponent 0 33 Low k Zone Apparent Tortuosity Factor Exponent 1 33 SOURCE ZONE CHARACTERISTICS Bulk Density of Transmissive Zone 1 70 g mL Source Zone Length Bulk Density of Low k Zone 1 70 g mL Source Zone Width 102100 Transverse Vertical Hydrodynamic Dispersivity 3 00E 04 f Restore 3 Distribution Coefficient mL g Source Loading Starts in Year 1962 format yyyy or Calculated R Source Removed in Year 2011 format yyyy Transmissive Zone Fraction of Organic Carbon 5 00E 04 1 GENERAL Low k Zone Fraction of Organic Carbon 1 50E 03 See Release Period Results for Organic Carbon Partitioning Coefficient 1 55E 02 L kg Year 2011 format yyyy Lateral Distance from Source 65 ft Depth into L
123. less Pp Bulk density of transmissive layer M L Ky Soil water partitioning coefficient L M foc Koc foc Fraction organic carbon of the transmissive layer unitless and Koc Organic carbon partitioning coefficient L M MATRIX DIFFUSION TOOLKIT v USER S MANUAL Y 109 APPENDIX A 2 DANDY SALE MODEL Appendix A 2 12 Transmissive Layer Total Mass Purpose Determine the transmissive layer total mass output in the Dandy Sale Model of the Matrix Diffusion Toolkit Given There is a finite amount of soluble organic constituents in the source zone in the dissolved sorbed and NAPL phases Assumptions The model makes the following assumptions 1 A source considered to be a thin pool is introduced at the contact between the two layers upgradient of x 0 2 A loading period occurs where there is a constant concentration of contaminants in the transmissive zone that drives contaminants into the low k zone 3 A release period occurs where the transmissive zone is assumed to have no concentration and an upper range estimate of release from the low k zone is generated 4 There is no degradation in either layer 5 Both layers are uniform homogeneous isotropic and infinite in the z direction perpendicular to groundwater flow 6 One dimensional 1 D advective transport in the transmissive layer parallel to the boundary of the layers is accompanied by transverse dispersion and diffusion 7 There is n
124. lso estimates the contaminant mass in both the low k and transmissive zones Mass is reported as aqueous phase sorbed phase and total aqueous plus sorbed PARAMETER NEXT STEP SAVE DATA Description Saves all the DSM model data DO NOT ADD ANY EXTENSIONS TO FILE NAME WHEN SAVING Note that this option does not save any edits performed on the graphs by the user To save such edits use the save function of Excel and save the entire Toolkit file PARAMETER EXPORT LOW k 2 D DATA Description Exports the 2 D low k aqueous sorbed and total concentration data into a text file for use in other programs The exported file follows the format MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 59 DSM RESULTS Lateral Distance from Source m or ft m or ft G N K 6 E E a D a Concentration mg L or mg g PARAMETER RETURN TO DSM DATA INPUT Returns to the DSM data input screen PARAMETER RETURN TO MAIN SCREEN Returns to the Toolkit Main Screen MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 60 DISSOLUTION MODEL DATA ENTRY NAPL DISSOLUTION MODEL DATA ENTRY Three important considerations regarding data input are 1 To see the example dataset in the input screen of the software click on the Paste Example button on the lower right portion of the input screen 2 Because the Toolkit is based on an Excel spreadsheet you have to click outside of the cell where you just entered data or hit Return before any of
125. lue directly Matrix Diffusion Toolkit Version 1 0 WB value calculated by Toolkit Do not enter data Site Location and ID Industrial Site Connecticut SYSTEM UNITS Transmissive Zone SI Units English Units T HYDROGEOLOGY Transmissive Zone Description Sand Gravel Transmissive Zone Effective Porosity 0 35 Low k Zone Description Silt silt k Low k Zone Total Porosity 0 43 Transmissive Zone Seepage Velocity 3 70E 01 x __Calculatev al TRANSPORT Key Constituent enter directly or choose from drop down list TCE Plume Loading Concentration Immediately Above Low k Zone in Vertical Plane Source During Loading Period 475 moh Le Molecular Diffusion Coefficient in Free Water 9 10E 10 m2 sec B Transmissive Zone Apparent Tortuosity Factor Exponent 9 Low k Zone Apparent Tortuosity Factor Exponent 33 4 SOURCE ZONE CHARACTERISTICS Bulk Density of Transmissive Zone g mL Source Zone Length 32 1 m Bulk Density of Low k Zone g mL Source Zone Width 39 3 m Transverse Vertical Hydrodynamic 1 00E 03 m Restore H Distribution Coefficient mL g Source Loading Starts in Year 1952 format yyyy or Calculated R Source Removed in Year 1996 format yyyy Transmissive Zone Fraction of Organic Carbon 3 80E 04 1 17 5 GENERAL Low k Zone Fraction of Organic Carbon 5 40E 04 1 18 See Release Period Results for Organic Carbon Partitioning Coefficient 9 3
126. ly to be an issue for most matrix diffusion modeling projects PARAMETER SEE CONC RESULTS Description Concentration in the transmissive zone from a monitoring well with a 10 foot screened interval This value is calculated using the mass discharge results as described above See the Intended Uses and Limitations Section for why the screen interval is hard wired to be 10 foot long and not a user input If you are sure all the mass discharge is being captured by a well with a different screened interval you can get the modeled concentration in this well by multiplying the model output by the ratio of the screens your screened interval 10 feet A lower range most likely value and an upper range for estimated outputs are provided based on the input source area concentrations The user may use the Log 2Linear button to see the results on a semi log plot MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 39 SRM RESULTS PARAMETER SEE MASS RESULTS Description Mass in the transmissive zone A lower range most likely value and an upper range for estimated outputs are provided based on the input source area concentrations The user may use the Log 2Linear button to see the results on a semi log plot PARAMETER RUN ADVANCED UNCERTAINTY ANALYSIS Description Uncertainty in parameter estimates is a key issue in estimating matrix diffusion effects The Toolkit provides two options for analyzing this uncertainty One opt
127. mental Security Testing and Certification Program ER 0530 Sale T C J A Zimbron and D S Dandy 2008b Effects of reduced contaminant loading on downgradient water quality in an idealized two layer granular porous media Journal of Contaminant Hydrology 102 2008 72 85 Seyedabbasi M A C J Newell D T Adamson and T C Sale 2012 Relative Contribution of DNAPL Dissolution and Matrix Diffusion to the Long Term Persistence of Chlorinated Solvent Source Zones J Cont Hydrology pp 69 81 DOI 10 1016 j jconhyd 2012 03 010 MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 74 REFERENCES Smith L and S W Wheatcraft 1993 Groundwater Flow in Handbook of Hydrology David Maidment Editor McGraw Hill New York New York Swiler L P and G D Wyss 2004 A User s Guide to Sandia s Latin Hypercube Sampling Software LHS Unix Library Standalone Version Sandia National Laboratories Albuquerque New Mexico TRRP 2008 Texas Risk Reduction Program RG 366 TRRP 19 Toxicity Factors and Chemical Physical Parameters June 2001 toxicity and physical chemical properties tables dated April 23 2008 http www tceq state tx us assets public remediation trrp trrptoxchph042308 xls USEPA 1997 Guiding Principles for Monte Carlo Analysis U S Environmental Protection Agency EPA 630 R 97 001 March 1997 Walton W C 1988 Practical Aspects of Groundwater Modeling National Water Well Association Worthington Ohio Wiedemeier T
128. meter For each run of the standard approach a random number is generated for the source concentration porosity apparent tortuosity factor exponent and retardation factor entered by the user This set of random inputs is then used to estimate concentration mass discharge and mass Repeating this procedure a large MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 16 MATRIX DIFFUSION TOOLKIT MODELS number of times yields a probability distribution from which statistical characteristics such as mean percentile and variance can be obtained The Toolkit employs 1000 iterations for the Monte Carlo analysis Guidelines for selecting key input parameters for the model are outlined in Uncertainty Analysis MATRIX DIFFUSION TOOLKIT v USER S MANUAL Y 17 SQUARE ROOT MODEL DATA ENTRY SRM Data Input Screen SRM Results Advanced Uncertainty Analysis Three important considerations regarding data input are 1 To see the example dataset in the input screen of the software click on the Paste Example button on the lower right portion of the input screen The example dataset used in the Toolkit is obtained from Chapman and Parker 2005 2 Because the Toolkit is based on an Excel spreadsheet you have to click outside of the cell where you just entered data or hit Return before any of the buttons will function Additionally REMOVING OR ADDING rows or columns in input screens may cause the program to crash 3 Parameters used in the model are
129. n the Toolkit input screen figure Typical Values 0 001 500 ft 0 001 152 m Source of Data Standard Method Enter the length your 2 modeled area Leave blank if you are only modeling one area Contour Map Method Use a contour map from the highest concentration MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 32 SRM DATA ENTRY period where groundwater samples were collected For example if concentrations have been decreasing use a concentration contour map from 1990 and not 2012 How to Enter Data Enter directly PARAMETER NEXT HIGHEST CONCENTRATION ZONE APPROXIMATE WIDTH W2 Units ft m Description Standard Method You can model two separate areas and the Toolkit will combine the diffusion processes This is the width of your second modeled area Contour Map Method Width of the second highest concentration contour line on a historical plume map between the upgradient and downgradient transects that represent your modeled area denoted by the blue box in the Toolkit input screen figure Typical Values 0 3 300 ft 0 1 000 m Source of Data Standard Method Modeled area width for this second of two subareas Contour Map Method Contour map should be from the highest concentration period where groundwater samples were collected For example if concentrations have been decreasing use a concentration contour map from 1990 and not 2012 How to Enter Data Enter directly PARAMETER CONCENTRATION OF CONTOUR L
130. ncentration of 1300 mg L also yielded a reasonable comparison without any input parameter value adjustments To determine how closely the Toolkit could match a declining source we took the estimated vs time curve from Chapman and Parker 2005 and assumed an average constant concentration of 475 mg L for 42 yrs However this did not show a good comparison with the observed concentrations A better match was obtained by assuming a constant 1100 mg L source active for 26 years Figure 1 9 Note that although only the source concentration and year in which the source was removed were used as calibration parameters for this evaluation other combinations of input parameters could be adjusted to yield similar results This shows that having actual data available for calibration improves the overall simulation results MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 123 CASE STUDY 1B INDUSTRIAL SITE CONNECTICUT DANDY SALE MODEL DSM Data Input Screen DATA_INPUT INSTRUCTIONS LIC Enter value directly Matrix Diffusion Toolkit Version 1 0 WB value calculated by Toolkit Do not enter data Site Location and ID Industrial Site Connecticut SYSTEM UNITS Transmissive Zone SI Units English Units T HYDROGEOLOGY Transmissive Zone Description Sand Gravel Transmissive Zone Effective Porosity 0 35 Low k Zone Description Silt silt k Low k Zone Total Porosity 0 43 Transmissive Zone Seepage Velocity 3 70E 01
131. ng Leco carbon analyzer a second core had foc values lt 0 005 for 10 samples and two samples with 0 00067 and 0 00084 gram per gram Values for foc using Walkley Black wet oxidation method were generally higher by a factor of MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 52 DSM DATA ENTRY 2 to 3 Values ranging from 0 to 0 078 have been reported for silts at the F W Warren site in Wyoming with a median value of 0 Source of Data The fraction organic carbon value should be measured if possible by collecting a sample of aquifer material from an uncontaminated saturated zone and performing a laboratory analysis e g ASTM Method 2974 87 or equivalent If unknown a default value of 0 002 should be used twice the typical default of 0 001 value used for transmissive systems How to Enter Data Enter directly PARAMETER ORGANIC CARBON PARTITIONING COEFFICIENT K lt Description Chemical specific partition coefficient between soil organic carbon and the aqueous phase Larger values indicate greater affinity of organic constituents for the organic carbon fraction of soil This value is chemical specific and can be found in chemical reference books Tetrachloroethene 155 mL g Benzene 66 mL g Trichloroethene 93 mL g Ethylbenzene 204 mL g cis 1 2 Dichloroethene 29 mL g Toluene 140 mL g Vinyl Chloride 11 mL g Xylene 240 mL g 1 1 1 Trichloroethane 110 mL g MTBE 14 mL g TRRP 2008 Note that there is a wide range of reported valu
132. nos Kettleman City area California U S Geol Survey open file report 182 p MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 73 REFERENCES Koerner R M Construction and Geotechnical Methods in Foundation Engineering McGraw Hill 1984 Lerman A 1979 Geochemical Processes in Water and Sediment Environments John Wiley and Sons New York Lovanh N Y Zhang R C Heathcote and P J J Alvarez 2000 Guidelines to Determine Site Specific Parameters for Modeling the Fate and Transport of Monoaromatic Hydrocarbons in Groundwater report submitted to the lowa Comprehensive Petroleum Underground Storage Tank Fund Board University of lowa lowa City lowa Mercer J W and R M Cohen 1990 A Review of Immiscible Fluids in the Subsurface Properties Models Characterization and Remediation Journal of Contaminant Hydrology 6 107 163 Millington R J and J P Quirk 1961 Permeability of Porous Media Nature 183 387 388 Newell C J J Gonzales and R K McLeod 1996 BIOSCREEN Natural Attenuation Decision Support System U S Environmental Protection Agency Center for Subsurface Modeling Support Ada OK EPA 600 R 96 087 Newell C J J A Connor and D L Rowan 2003 Groundwater Remediation Strategies Guide American Petroleum Institute Publication Number 4730 Washington D C December 2003 Newell C J S K Farhat D T Adamson and B B Looney 2011 Contaminant Plume Classification System Based on Mas
133. nt transect line perpendicular to groundwater flow and an upgradient transect line perpendicular to groundwater flow to define the area you want to assess using the Toolkit Here are three examples where you need to enter the length and width of the areas representing the source black box and blue box in the drawing MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 27 SRM DATA ENTRY Type of Problem to be Analyzed Using the Black Box in Blue Box in Toolkit Drawing Drawing To see matrix diffusion impacts in a source The Black Box is The Blue Box is zone drawn around the drawn around the highest contour in second highest the source area contour in the source area Note You want to use a contour map with the highest concentrations measured at the site to represent higher historical concentrations To see matrix diffusion impacts in a The Black Box is The Blue Box is downgradient plume drawn around the drawn around the highest contour second highest downgradient of the contour source area downgradient of the source area To see matrix diffusion impacts downgradient The Black Box is The Blue Box is of a PRB drawn around the drawn around the highest contour second highest downgradient of the contour PRB downgradient of the PRB The width of the box is the width of the PRB Step 5 2 Standard Method Enter your own length and width for the first modeled area L and W1 Contour Lin
134. o longitudinal dispersion in the transmissive layer 8 1 D transverse diffusion transport occurs in the low k layer 9 Retardation of contaminants in both layers is based on instantaneous equilibrium between aqueous and sorbed phases Summary Total mass in the transmissive layer at any time t can be calculated as Miot t z Maq t M t where Mu Total mass in the transmissive layer at any time t M M t Aqueous phase mass in the transmissive layer at any time t M calculated using Appendix A 2 10 and M t Sorbed phase mass in the transmissive layer at any time t M calculated using Appendix A 2 11 MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 110 APPENDIX A 3 PROBABILITY DISTRIBUTIONS This section describes in greater detail the probability distributions employed in the Monte Carlo analysis The Matrix Diffusion Toolkit offers the user three distribution options normal lognormal and uniform A 3 1 Normal Distributions Normal distributions are defined by the density function eu 20 o x oo 1 Ho elon where o is the standard deviation and u the mean of the distribution The Toolkit assumes that the values entered in the Input Data and Grid screen are the means The uncertainty analysis requires the user to specify a o as a percentage of the mean A 3 2 Lognormal Distributions A lognormal distribution is a distribution whose logarithms are normally distributed The lognormal densit
135. o the user may have to rely on scientific engineering judgment to use the Monte Carlo analysis e The Monte Carlo analysis cannot account for plume data that are not part of the monitoring system Actual mass discharge concentration and mass values can be outside the reported range of mass flux values from the Monte Carlo analysis for example if new data show high concentration zones that were not captured by the original monitoring network e The Square Root model assumes an unimpeded release during the release period In other words for purposes of calculating the rate at which contaminants diffuse out of the low k zones the model assumes there is no concentration in the transmissive zone Because diffusion from a low k zone is a relatively weak force compared to active DNAPL sources and because the model assumes an instantaneous switch from loading to release period this assumption should not prevent the model from providing useful order of magnitude type information e The Square Root model assumes that the loading of the low k zone is a horizontal area directly over the low k zone This assumption can be applied to source zones such as ones that contained DNAPL pools or to downgradient parts of the plume where a high concentration aqueous phase plume provided the loading to the low k zone e The Dandy Sale model basically assumes the source zone is a vertical plane and only estimates the effect of matrix diffusion downgradient
136. od comparison to soil core concentrations 65 ft downgradient of the source Note that although for this evaluation only the seepage velocity source concentration low k fraction organic carbon and the diffusion coefficient were used as calibration parameters there could be other combinations of input parameters could be adjusted to yield similar or better results After working to match the soil core data the model now can be used to estimate future concentrations and mass discharge in the low k zone at the site MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 147 CASE STUDY 3 DSM Data Input Screen Matrix Diffusion Toolkit DANDY SALE MODEL Version 1 0 Site Location and ID Jacksonville Florida SYSTEM UNITS SI Units HYDROGEOLOGY Transmissive Zone Description Transmissive Zone Effective Porosity English Units Low k Zone Description Low k Zone Total Porosity Transmissive Zone Seepage Velocity TRANSPORT Key Constituent enter directly or choose from drop down list Plume Loading Concentration Immediately Above Low k Zone in Vertical Plane Source During Loading Period Molecular Diffusion Coefficient in Free Water Transmissive Zone Apparent Tortuosity Factor Exponent Low k Zone Apparent Tortuosity Factor Exponent Bulk Density of Transmissive Zone Bulk Density of Low k Zone Distribution Coefficient or Transmissive Zone Fraction of Organic Carbon Low k Zone Fraction of Organic Carbon
137. ollowing assumptions 1 A vertical plane source at X 0 is assumed This vertical plane has concentrations that decrease exponentially in the vertical direction the farther MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 84 APPENDIX A 2 DANDY SALE MODEL one gets from the bottom of the transmissive zone This exponential pattern is defined using the length of the source materials L upgradient of the vertical plane source and other factors using Equation 3 it assumes vertical dispersion accounts for the vertical pattern 2 A loading period occurs where there is a constant concentration of contaminants in the transmissive zone that drives contaminants into the low k zone 3 A release period occurs where the transmissive zone is assumed to have no concentration and an upper range estimate of release from the low k zone is generated 4 There is no degradation in either layer 5 Both layers are uniform homogeneous isotropic and infinite in the z direction perpendicular to groundwater flow 6 One dimensional 1 D advective transport in the transmissive layer parallel to the boundary of the layers is accompanied by transverse dispersion and diffusion 7 There is no longitudinal dispersion in the transmissive layer 8 1 D transverse diffusion transport occurs in the low k layer 9 Retardation of contaminants in both layers is based on instantaneous equilibrium between aqueous and sorbed phases Summary Active Source While
138. or the Toolkit we assume that the transverse vertical hydrodynamic dispersivity is relatively small no more than 0 001 meters for two reasons 1 there is a new low dispersion paradigm emerging in the solute transport field and 2 since the Toolkit is calculating concentration from a horizontal mass flux equation we have to assume a plume never disperses more than 10 feet above the low k transmissive zone contact we assume a 10 foot monitoring well is used to determine groundwater concentrations in the model MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 55 DSM DATA ENTRY Typical Values For this model o x 0 0004 ft 0 001 m Source of Data Typically estimated using empirical relationships How to Enter Data Enter directly The Toolkit automatically assigns a maximum value of 0 0004 ft 0 001 m This value can be overwritten Use the Restore button to restore the Toolkit calculated value Step 5 General PARAMETER LATERAL DISTANCE FROM SOURCE x Description Lateral distance from source for displaying matrix diffusion results Note Due to computational limitations all transmissive zone solutions show increasing numerical imprecision for lengths greater than 4921 ft 1500 m Therefore we recommend limiting the lateral distance from the source to S 4921 ft 1500 m for any model runs involving transmissive zone solutions Appendix A 2 7 Equation 1 How to Enter Data Enter directly PARAMETER DEPTH INTO LOW K ZON
139. ost cases A value of 1 7 g mL is used frequently for unconsolidated media Representative values in g mL for specific geologic media are shown below Lovanh et al 2000 derived from Domenico and Schwartz 1990 Clay 1 0 2 4 Loess 0 75 1 6 Sands ne 1 6 2 68 Shale 1 54 3 17 Limes ne 1 74 2 79 Granite 2 24 2 46 Basalt 2 2 7 Medium Sand 1 34 1 81 Koerner 1984 reports these values in g mL for unit weight for saturated soils note no dry bulk density values are reported for these materials Glacial till very mixed grain 2 32 Soft glacial clay 1 77 Stiff glacial clay 2 07 Soft slightly organic clay 1 58 Soft very organic clay 1 43 Soft bentonite 1 27 Source of Data Either from an analysis of soil samples at a geotechnical lab or more commonly application of estimated values such as 1 7 g mL How to Enter Data Enter directly PARAMETER LOW k ZONE FRACTION ORGANIC CARBON foc Description Fraction of the aquifer material comprised of natural organic carbon in uncontaminated areas More natural organic carbon means higher adsorption of organic constituents on the aquifer matrix Typical Values Although based on limited data 0 0002 0 10 for low K zones is a likely range But some sites may be higher or lower Examples At the Moffatt Field site the foc of the clay fraction is about 0 0066 Roberts et al 1990 Domenico and Schwartz 1990 report these values silt Wildwood Ontario 0 00102 f
140. ous data above as the geometric mean of highest contour line and second highest contour concentration This value can be overwritten How to Enter Data Enter directly or let the Toolkit calculate it Note that if overwritten the Toolkit calculated value in the blue cell can be replaced by pressing the Restore button PARAMETER UNCERTAINTY IN PLUME CONCENTRATION ESTIMATIONS Description Users should make a realistic estimate on how much uncertainty is associated with the concentration estimation being modeled The main point of this parameter and the software to some extent is that there is a high level uncertainty in any source concentration estimation A value of factor of 10 is typically used Note that if a value of 1 is used then only the most likely estimate line will be shown on the graphs How to Enter Data Enter directly Step 6 General PARAMETER SOURCE LOADING STARTS IN YEAR Description Year source loading started This is estimated from site historical records and is almost always from the 1950s 1960s 1970s or early 1980s If the release was over a long period of time usually it is better to enter the earliest year This can be used as a calibration parameter see the beginning of this section How to Enter Data Enter directly MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 34 SRM DATA ENTRY PARAMETER SOURCE REMOVED IN YEAR Description Year source was removed This is either 1 the
141. ow k Zone z 16 5 ft Next Step New Site Clear Data HELP Show Graph Save Data Load Data Show Previous Results Return to Model Selection Return to Main Screen Figure 3 3 DSM Input Parameters Calibrated MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 149 CASE STUDY 3 FORMER DRY CLEANER FLORIDA DANDY SALE MODEL Low k Zone Aqueous Concentration mg L for Year 2011 Lateral Distance from Source ft 19 50 22 75 26 00 29 25 32 50 35 75 39 00 45 50 52 00 55 25 58 50 61 75 Se O o E f x e a o 2 s z 2 o a oi m m a mg Low k Zone Aqueous Mass ERU kg Sorbed Mass kg Total Mass kg m0 13 m13 26 m26 39 m39 52 m52 65 m65 77 m77 90 m90 103 m103 116 m116 129 m129 142 9 See Trans Zone Aqueous Conc See Low k Aq C Dist EE See Trans Zone Mass Discharge See Low k Aq Conc vs Depth See 2 D Low k Sorbed Conc See 2 D Low k Total Conc See Trans Zone Sorbed Conc See Trans Zone Total Conc Figure 3 4 DSM Output for OU3 3 Calibrated MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y Export Low k 2 D Data HELP Return to DSM Data Input 15
142. ow k zone How to Enter Data Choose from drop down list or enter directly PARAMETER LOW k ZONE TOTAL POROSITY 6 Description Dimensionless ratio of the volume of voids to the bulk volume of the surface soil column matrix but excluding secondary porosity fractures solution cavities etc Total porosity is the ratio of all voids including non connected voids to the bulk volume of the aquifer matrix Effective porosity and any porosity data with secondary porosity information should not be used Typical Values The model input screen has these default values Clay 0 47 mid range of values below Silt 0 48 mid range of values below Sandstone shale 0 10 Pankow and Cherry 1996 Table 12 2 Fractured Sandstone 0 08 Pankow and Cherry 1996 Table 12 2 Granite 0 006 Pankow and Cherry 1996 Table 12 2 Values for total porosity from Domenico and Schwartz 1990 in part from Davis 1969 and Johnson and Morris 1962 SEDIMENTARY Porosity Gravel coarse 0 24 0 36 Gravel fine 0 25 0 38 Sand course 0 31 0 46 Sand fine 0 26 0 53 Silt 0 34 0 61 Clay 0 34 0 60 SEDIMENTARY ROCKS Sandstone 0 05 Siltstone 0 21 Shale 0 0 CRYSTALLINE ROCKS Dense crystalline rocks 0 0 05 0 30 0 41 0 1 Koerner 1984 reports these values for unit weight for saturated soils note no dry bulk density values are reported for these materials Glacial till very mixed grain 0 20 Soft glacial clay 0 57 Stiff glacial
143. plying the model output by the ratio of the screens your screened interval 10 feet MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 99 APPENDIX A 2 DANDY SALE MODEL Contaminant flux at the layer interface can be derived by obtaining the derivative of concentration in the low k zone with respect to z at z 0 Active Source Contaminant mass flux across the low k transmissive layer boundary at any lateral distance from the source and time can be obtained from b t TE n Ame m Yal v i with b o D D Va R R and y D Va De 5 D nD 6 V V R 7 R 1 S 8 R 1 DE 9 pec 09 where J x t Contaminant flux at distance x and time t M T Mean plume loading concentration above the low k layer during the charging period M L a Coefficient of transverse hydrodynamic dispersion L b Source characteristic 1 L D Effective transverse diffusion coefficient in the low k layer L T De Effective molecular diffusion coefficient in the transmissive layer L T Do Molecular diffusion coefficient in free water L T MATRIX DIFFUSION TOOLKIT v USER S MANUAL Y 100 APPENDIX A 2 DANDY SALE MODEL D Effective transverse diffusion coefficient in the transmissive layer L T foc Fraction organic carbon of the transmissive layer unitless P Fraction organic carbon of the low k layer unitless Koc Organic carbon partitioning coefficient L M L Source
144. ptions The Toolkit uses a simplified conceptual model of a two layer aquifer system a transmissive layer and a low k layer where there are two different time periods 1 A loading period where there is a constant concentration of contaminants in the transmissive zone that drives contaminants into the low k zone 2 A release period where the transmissive zone is assumed to have no concentration and an upper range estimate of release from the low k zone is generated 3 The low k zone is at least 1 meter thick 4 There is no degradation in the low k zone 5 Mass discharge is occurring at a well with a 10 ft screened interval Summary At any time t the concentration of contaminants in a transmissive zone can be estimated using the equation Gp V HW Where C t Plume concentration in the transmissive zone at time t M L Mg Mass discharge from the low k layer into the transmissive layer L T Va Darcy velocity of the transmissive compartment L T H Screened interval of the hypothetical well L and W Width of the modeled area L MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 81 APPENDIX A 1 SQUARE ROOT MODEL Appendix A 1 3 Estimation of Mass in Transmissive Zone Purpose Estimate the transmissive zone mass output of the Square Root Model of the Matrix Diffusion Toolkit Given There is a finite amount of soluble organic constituents in the source zone in the dissolved sorbed and NAPL phases A
145. rea of interest screened close to the low k unit being modeled To match model output the actual monitoring wells in the field should have screens long enough to capture any of the contaminant mass diffusing off the low k zone In other words if possible you should use data from wells with 5 to 10 foot screened intervals and not from shorter screened intervals How to Enter Data Enter directly PARAMETER FIELD DATA FOR COMPARISON MASS DISCHARGE Description Mass discharge measurements in transmissive zone and or low k zone area of interest These data are displayed with model results in the Next Step Show Graph option Low k zone mass discharge should be entered as negative values Typical Values 0 001 10 000 g d Source of Data Transects of wells located in the area of interest pumping well data or flux meters How to Enter Data Enter directly MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 36 SRM DATA ENTRY PARAMETER FIELD DATA COMPARISON MASS Units Description measurements in transmissive zone area of interest These data are displayed with model results in the Next Step Show Graph option Typical Values 0 10 100 000 kg Source of Data Soil samples located in the area of interest pumping well data calculated from groundwater data and saturated soil constituent concentration data or estimated from NAPL relationships Software tools such as SourceDK Farhat et al 2004 can be used to
146. reen figure Typical Values 0 3300 ft 0 1000 m Source of Data Standard Method Modeled area width for this first of two subareas Contour Map Method Contour map should be from the highest concentration period where groundwater samples were collected For example if concentrations have been decreasing use a concentration contour map from 1990 and not 2012 How to Enter Data Enter directly PARAMETER HIGHEST HISTORICAL CONCENTRATION IN BLACK BOX Cs Description Standard Method Leave this blank and just enter the historical loading concentration for the first modeled area in Representative Concentrations Contour Map Method The highest maximum observed concentration in the modeled area black box area defined by the length and width above Typical Values 0 0001 20 000 mg L Source of Data Standard Method Not needed Contour Map Method Use a contour map from the highest concentration period where groundwater samples were collected For example if concentrations have been decreasing use a concentration contour map from 1990 and not 2012 How to Enter Data Enter directly PARAMETER CONCENTRATION OF CONTOUR LINE IN BLACK BOX Description Standard Method Leave this blank and just enter the historical loading concentration for the first modeled area in Representative Concentrations Contour Map Method Concentration of contour line represented by the black box in the Toolkit input screen figure Typi
147. refer MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 53 DSM DATA ENTRY Transmissive Zone cating tio AT Low k Zone vertical plane source This is conceptualized in Sale et al 2008b as a DNAPL pool upgradient of the modeled zone the Dandy Sale Model simulates matrix diffusion in the downgradient plume see figure below Square Root Dandy Sale sorbed DNAPL constituents Typical Values 10 500 ft 3 152 m Source of Data To determine source length across the site draw a line parallel to the direction of groundwater flow in what is considered to be the high concentration source area The DSM source length is not a highly sensitive parameter in the model if unsure of which value to use enter about 100 ft How to Enter Data Enter directly PARAMETER SOURCE ZONE WIDTH W The estimated width of the source zone perpendicular to the groundwater flow Typical Values 0 500 ft 0 152 m Source of Data To determine source width across the site draw a line perpendicular to the direction of groundwater flow in what is considered to be the high concentration source area MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 54 DSM DATA ENTRY gt Es d How to Enter Data Enter directly PARAMETER SOURCE LOADING STARTS IN YEAR Description Year source loading started Estimated from site historical records and is usually from the 1950s 1960s 1970s or early1980s If the release was over a long period of
148. rom Oconee River sediment coarse silt 0 029 medium silt 0 02 fine silt 0 0226 Chapman and Parker 2005 report a foc of glaciolacustrine aquitard composed of varved silts and clays 0 0024 to 0 00104 with an average of 0 00054 Adamson 2012 reports foc 0 001 for a clay layer in Jacksonville Florida and foc values for silts at the MMR site in Massachusetts ranging from lt 0 0005 to 0 0022 median value 0 0014 for one core using Leco carbon analyzer a second core had foc values lt 0 005 for 10 samples and two samples with 0 00067 and 0 00084 gram per gram Values for foc using Walkley Black wet oxidation method were generally higher by a factor of 2 to 3 Values ranging from 0 to 0 078 have been reported for silts at the F W Warren site in Wyoming with a median value of 0 Source of Data The fraction organic carbon value should be measured if possible by collecting a sample of aquifer material from an uncontaminated saturated zone and performing a laboratory analysis e g ASTM Method 2974 87 or equivalent If unknown a default value of 0 002 should be used twice the typical default of 0 001 value used for transmissive systems How to Enter Data Enter directly MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 25 SRM DATA ENTRY PARAMETER ORGANIC CARBON PARTITIONING COEFFICIENT Koc Description Chemical specific partition coefficient between soil organic carbon and the aqueous phase Larger values indicate gre
149. s section page 4 the DSM model assumes a two layer system with one interface an infinitely thick low k zone and instantaneous flushing of the transmissive zone instantly changing from the loading period to the release period The tank experiment had four very thin low k zones ranging from 0 03 to 0 2 m thick compared to a theoretical contaminant penetration depth into an infinite low k zone of 0 25 m Despite these differences from the assumed configuration of the DSM end results show the model was able to match actual data from the tank within an order of magnitude MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 143 DSM Data Input Screen Matrix Diffusion Toolkit Site Location and ID Tank Stud 1 SYSTEM UNITS SI Units English Units HYDROGEOLOGY Transmissive Zone Description Transmissive Zone Effective Porosity Low k Zone Description Low k Zone Total Porosity Transmissive Zone Seepage Velocity TRANSPORT Key Constituent enter directly or choose from drop down list Plume Loading Concentration Immediately Above Low k Zone p Vertical Plane Source During Loading Period Molecular Diffusion Coefficient in Free Water Transmissive Zone Apparent Tortuosity Factor Exponent Low k Zone Apparent Tortuosity Factor Exponent Bulk Density of Transmissive Zone Bulk Density of Low k Zone Distribution Coefficient or Transmissive Zone Fraction of Organic Carbon Low k Zone Fraction of Organic Carbon Organic Carbon Partit
150. s Discharge Ground Water 49 no doi 10 1111 j 1745 6584 2010 00793 x Pankow J F and J A Cherry 1996 Dense Chlorinated Solvents and other DNAPLs in Groundwater Waterloo Press Portland Oregon Pantazidou M and K Liu 2008 DNAPL Distribution in the Source Zone Effect of Soil Structure and Uncertainty Reduction with Increased Sampling Density Journal of Contaminant Hydrology 96 169 186 Parker B L R W Gillham and J A Cherry 1994 Diffusive Disappearance of Immiscible Phase Organic Liquids in Fractured Geologic Media Groundwater 32 5 805 820 Payne F C J A Quinnan and S T Potter 2008 Remediation Hydraulics CRC Press Boca Raton Florida Roberts P V G D Hopkins D M Mackay and L Semprini 1990 Field Evaluation of In Situ Biodegradation of Chlorinated Ethenes 1 Methodology and Field Site Characterization Ground Water 28 4 591 604 Rong Y R F Wang and R Chou 1998 Monte Carlo Simulation for a Groundwater Mixing Model in Soil Remediation of Tetrachloroethylene Journal of Soil Contamination 7 1 87 102 Sale T C 1998 Interphase Mass Transfer from Single Component DNAPLs Ph D Dissertation Department of Chemical and Bioresource Engineering Colorado State University Fort Collins Colorado Sale T C C J Newell H Stroo R Hinchee and P J Johnson 2008a Frequently Asked Questions Regarding Management of Chlorinated Solvent in Soils and Groundwater Developed for the Environ
151. s a multiplier to give a range around the most likely value The Square Root Model also utilizes a Monte Carlo type approach to analyze uncertainty in the actual concentration porosity apparent tortuosity factor exponent and retardation factor measurements With this tool groundwater practitioners can estimate the accuracy of the hydrologic measurements that are being used for the matrix diffusion calculation The Dandy Sale Model is likely to have the same level of accuracy as the Square Root Model Because of the complexity of this model we currently don t show the order of magnitude results in the model output Of course with more field data especially sampling results from the low k zone the accuracy of the modeling results will increase MATRIX DIFFUSION TOOLKIT v USER S MANUAL Y 14 MATRIX DIFFUSION TOOLKIT MODELS Two models are utilized in the Toolkit the Square Root Model and the Dandy Sale Model Square Root Dandy Sale TA nee of asain and sorbed DNAPL constituents Square Root Model SRM Building on work originally performed by Drs Beth Parker and John Cherry and modified by Dr Tom Sale the Square Root Model SRM provides planning level estimates of the mass discharge in units of grams per day caused by release from a low k diffusion dominated unit typically silt or clay into a high permeability advection dominated unit typically sand or gravel The Toolkit also estimates
152. s for the low k layer unitless Bulk density of transmissive layer M L Bulk density of low k layer M L Time since source was introduced T Seepage velocity in the transmissive layer L T Contaminant transport velocity in the transmissive layer L T Lateral distance from source edge L Dummy integration variable and Depth of the low k layer L E D 9 So DDD H G Q o e KE EE on b lt 2 N dX Exhausted Source Once the source is exhausted the low k aqueous concentration can be calculated at any time t using Sale et al 2008b MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 86 APPENDIX A 2 DANDY SALE MODEL I I b t _ ES x GZ LE _ 1 xz GEI Sdk a b b CUZ treo E T Je E erfc L NI 11 where t is the source persistence time e the time in which the source is active and 1 x Z t t Z is defined as z erfc 2 ayaa S y Pr D z t v ee RI Ve m Y Q2zr2 2 t t Ve D R x y ei sr bye x Soy exp z KEE du NET 9 t t Zil I x z t 1 erfc 12 Boundary Conditions C x z 0 0 z 0 13 C x z 0 0 lt z lt 0 14 C xz 20 t 20 15 C x z gt o t 0 16 ac ma C nD x 0 t nD 5 x 0 t 17 The source introduced in the transmissive layer at x 0 is modeled as C 0 z t Coe 1 H t 1 z2 0 18 where C is the aqueous concentrat
153. samples or from literature values How to Enter Data Enter directly PARAMETER NAPL SATURATION Sy An estimate of the fraction of the pore space filled with NAPL Typical Values 0 0 30 For a discussion of NAPL saturation at solvent sites see Pankow and Cherry 1996 For a detailed discussion of solvents and fuels see Mercer and Cohen 1990 For a brief summary see Chapter 2 of Wiedemeier et al 1999 Source of Data This value can be measured by analyzing soil samples Without site specific measurements the uncertainty in the estimates will likely be an order of magnitude or greater How to Enter Data Enter directly MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 67 DISSOLUTION MODEL RESULTS NAPL Dissolution Model Results PARAMETER MASS FLUX DUE TO DIFFUSION FROM TOP OF NAPL POOL Mass flux due to diffusion from the top of the NAPL pool PARAMETER NUMBER OF YEARS FOR DISSOLUTION OF NAPL PLUME The time required for the dissolution of the NAPL plume PARAMETER NEXT STEP SAVE DATA Description Saves all the NAPL Dissolution model data DO NOT ADD ANY EXTENSIONS TO FILE NAME WHEN SAVING PARAMETER NEW SITE CLEAR DATA Description Clears ALL data related to the NAPL Dissolution model in the Toolkit memory banks Use this button to start a new project PARAMETER PASTE EXAMPLE Description Clears ALL data related to the NAPL Dissolution model in the Toolkit memory banks and pastes an example dataset PARAMETER LOAD
154. screened interval The 10 foot screened interval was selected because at an actual field site contamination diffusing from a low k zone might spread vertically above a 1 foot screen It was thought to be very unlikely that there would be more than 10 feet of vertical spreading in the transmissive zone Bottom line the 10 foot screened interval is hard wired into the model and cannot be changed by the user If you are sure all the mass discharge is being captured by a well with a MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 58 DSM RESULTS different screened interval you can get the simulated concentration in this well by multiplying the model output by the ratio of the screens your screened interval 10 feet The user may use the Log 2Linear button to see the results on a semi log plot PARAMETER SEE TRANS ZONE MASS DISCHARGE Description Mass discharge vs distance in the transmissive zone The user may use the Log 2Linear button to see the results on a semi log plot PARAMETER SEE TRANS ZONE SORBED CONC Description Sorbed phase concentration vs distance in the transmissive zone The user may use the Log 2Linear button to see the results on a semi log plot PARAMETER SEE TRANS ZONE TOTAL CONC Description Total concentration vs distance in the transmissive zone The user may use the Log 2Linear button to see the results on a semi log plot PARAMETER MASS RESULTS Description The Toolkit a
155. smissive zone The user may use the Log 2Linear button to see the results on a semi log plot PARAMETER SEE CONC RESULTS Description The 5 percentile median and 95 percentile for concentration from a well with a 10 foot screen in the transmissive zone based on the user s choice of interpolation method and uncertainty in the input variables as defined by their probability distributions means variances and ranges The user may use the Log lt gt Linear button to see the results on a semi log plot PARAMETER SEE MASS RESULTS Description The 5 percentile median and 95 percentile for mass in the transmissive zone based on the user s choice of interpolation method and uncertainty in the input variables as defined by their probability distributions means variances and ranges The user may use the Log 2Linear button to see the results on a semi log plot PARAMETER SAVE DATA Description Saves all the SRM model data DO NOT ADD ANY EXTENSIONS TO FILE NAME WHEN SAVING Note that this option does not save any edits performed on the graphs by the user To save such edits use the save function of Excel and save the entire Toolkit file PARAMETER RETURN TO SRM RESULTS Returns to the SRM Model Results screen PARAMETER RETURN TO SRM DATA INPUT Returns to the SRM data input screen MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 42 SRM ADVANCED UNCERTAINTY ANALYSIS PARAMETER EXPORT UNCERTAINTY DATA Descr
156. ssumptions The Toolkit uses a simplified conceptual model of a two layer aquifer system a transmissive layer and a low k layer where there are two different time periods 1 A loading period where there is a constant concentration of contaminants in the transmissive zone that drives contaminants into the low k zone 2 A release period where the transmissive zone is assumed to have no concentration and an upper range estimate of release from the low k zone is generated 3 The low k zone is at least 1 meter thick 4 There is no degradation in the low k zone Summary At any time t mass discharge into the low k layer underlying the pool can be estimated using the equation RD RD M_ t t C A e Where Ma Mass discharge M T t Time since source was introduced T t Time at which source was removed from the high permeability compartment T Porosity of low k zone unitless C Mean plume concentration above the low k compartment M L A lt Area of silt compartment beneath the plume L R Retardation factor for low k compartment unitless and D Effective aqueous phase diffusion coefficient in the low k compartment L T This can be estimated as D 9 D where p is the apparent tortuosity factor exponent unitless and D the molecular diffusion coefficient in free water L T Integrating this equation for mass yields Seyedabbasi et al 2012 MATRIX DIFFUSION TOOLKIT Y U
157. ssumptions The model makes the following assumptions 1 A source considered to be a thin pool is introduced at the contact between the two layers upgradient of x 0 2 A loading period occurs where there is a constant concentration of contaminants in the transmissive zone that drives contaminants into the low k zone 3 A release period occurs where the transmissive zone is assumed to have no concentration and an upper range estimate of release from the low k zone is generated 4 There is no degradation in either layer 5 Both layers are uniform homogeneous isotropic and infinite in the z direction perpendicular to groundwater flow 6 One dimensional 1 D advective transport in the transmissive layer parallel to the boundary of the layers is accompanied by transverse dispersion and diffusion 7 There is no longitudinal dispersion in the transmissive layer 8 1 D transverse diffusion transport occurs in the low k layer 9 Retardation of contaminants in both layers is based on instantaneous equilibrium between aqueous and sorbed phases Summary Since the medium is saturated with water the water content equals the porosity Consequently the total concentration mass of the constituent per unit bulk volume can be obtained using n ppKa Ctota o t Cweu Db 1 where Crota X t Total concentration at lateral distance x and any time t M M Cell Well concentration at lateral distance x and any time t MI Calc
158. stance away from the source edge In this model z is designated as the vertical depth from the source in the low k layer and z the height in the transmissive zone Transmissive Zone Low k Zone Results Calculated Here Figure A 2 7 1 The two layer scenario conceptual model Top Active Source Bottom Depleted Source The model makes the following assumptions 1 A source considered to be a thin pool is introduced at the contact between the two layers upgradient of x 0 MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 98 APPENDIX A 2 DANDY SALE MODEL 2 A loading period occurs where there is a constant concentration of contaminants in the transmissive zone that drives contaminants into the low k zone 3 A release period occurs where the transmissive zone is assumed to have no concentration and an upper range estimate of release from the low k zone is generated 4 There is no degradation in either layer 5 Both layers are uniform homogeneous isotropic and infinite in the z direction perpendicular to groundwater flow 6 One dimensional 1 D advective transport in the transmissive layer parallel to the boundary of the layers is accompanied by transverse dispersion and diffusion 7 There is no longitudinal dispersion in the transmissive layer 8 1 D transverse diffusion transport occurs in the low k layer 9 Retardation of contaminants in both layers is based on instantaneous equilibrium between aqu
159. storical groundwater concentrations in the early period of a plume s life cycle Data Source 2 Highest Observed Concentration More commonly good Data Source 1 information will not be available In that case we recommend using the highest observed concentration from a groundwater monitoring point in the modeled area the two boxes and a groundwater concentration contour map from the period with the highest observed concentrations from the monitoring network This is typically the oldest concentration contour map available While not perfect this method is based on real data and represents observed loading concentrations in the modeled area This value is a key parameter that can be changed during the calibration process to increase or decrease the simulated mass discharge or concentration to better match field data see the beginning of this section MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 49 DSM DATA ENTRY PARAMETER MOLECULAR DIFFUSION COEFFICIENT IN FREE WATER D eH A factor of proportionality representing the amount of substance diffusing Description WU Pur across a unit area through a unit concentration gradient in unit time Typical Values Benzene 9 8E 06 cm s Tetrachloroethene 8 2E 06 cm s Ethylbenzene 7 8E 06 cm s Trichloroethene 9 1E 06 cm s Toluene 8 6 06 cm s cis 1 2 Dichloroethene 1 1E 05 cm s Xylene 8 5E 06 cm s Vinyl Chloride 1 2E 05 cm s MTBE 9 4E 05 cm s 1 1 1 Trichloroethane 8 8E
160. t in free water Apparent tortuosity factor exponent Retardation factor Plume Characteristics High concentration zone Approximate length Approximate width Highest concentration in black box Concentration of contour line in black box Representative concentration Next highest concentration zone Approximate length Approximate width Concentration of contour line in blue box Representative concentration Uncertainty in plume concentration estimations Field Data for Effluent bromide and fluorescein concentrations Comparison Sand tank construction Estimated sand tank study Experimental sand tank study Sand tank Literature sand tank study clay 0 60 1 51E 4 cm sec bromide fluorescein bromide 2 01E 9 m sec fluorescein 5 5E 10 m sec 1 US Literature sand tank study Assumed sand tank study Based on area of lengths of clay layers in sand tank and 0 71 m m width of tank 0 03 m 1 mg L 1 mg L Initial 1 mg L Calibrated 1 88 mg L fluorescein 0 71 m 0 03 m 1 mg L Initial 1 mg L Calibrated 1 88 mg L fluorescein 10 factor of e See Figures 2 4 and 2 7 General Source loading starts in year 2006 Sand tank study Source removed in year 24 days Sand tank stud MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 134 CASE STUDY 2A SAND TANK STUDY
161. ta than the original estimate of loading concentration 475 mg L over 42 years MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 132 CASE STUDY 2 SAND TANK STUDY Overview The Toolkit was used to estimate the effects of diffusion into and from low k zones for tracers bromide and fluorescein in a sand tank The sand tank study is described in detail in Chapman et al 2012 For this analysis both the SRM Case Study 2A and the DSM Case Study 2B were applied as follows e Step 1 Initial values of all parameters obtained from either Chapman et al 2012 or default Toolkit parameters were entered into the Toolkit e Step 2 Toolkit outputs were compared to observed tracer concentrations This step was critical in determining how well Toolkit default parameters predicted actual field conditions e Step 3 Input parameters were adjusted as needed to improve the comparison with observed tracer concentrations Figure 2 1 Sand Tank Configuration Based on Chapman et al 2012 Figure 1 Darker shaded areas A B C and D represent low k bentonite zones lying in transmissive sandy zones MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 133 CASE STUDY 2A SAND TANK STUDY SQUARE ROOT MODEL A Square Root Model SRM Input Data Data Type Source of Data Hydrogeology Low k zone description Low k zone porosity Darcy velocity Transport Low k Zone Key constituent Molecular diffusion coefficien
162. te groundwater concentrations in the low k zone in both the source and plume areas Toolkit input and output for the source zone analysis are shown on Figures 1 4 and 1 5 A comparison of the Toolkit output with observed values is shown on Figure 1 6 Toolkit input and output for the plume zone analysis are shown on Figures 1 7 through 1 11 A comparison of the Toolkit output with observed values is shown on Figure 1 12 Site hydrogeological data was entered in Section 2 transport parameters in Section 3 source zone characteristics in Section 4 and desired output information in Section 5 Site specific values as documented by Chapman and Parker 2005 were available for all parameters except molecular diffusion coefficient in free water apparent tortuosity factor exponent organic carbon partitioning coefficient and coefficient of transverse hydrodynamic coefficient For these Toolkit default values were used Additionally the Toolkit default value for the organic carbon partitioning coefficient was also used A sheet pile enclosure was installed in 1994 around the DNAPL area Figure 1 1 o Forthe source zone analysis field comparison data were collected inside the sheet pile enclosure in 1997 therefore for this analysis the source was assumed to be active in 1997 o Forthe plume zone analysis to account for the travel of contaminated groundwater present at the time of the sheet pile an effective sourc
163. th Source Zone Width Transverse Vertical Hydrodynamic Dispersivity Source Loading Starts in Year Source Removed in Year 32 1 m 39 3 m 1 00E 03 m 1952 format yyyy 1978 format yyyy Restore zl See Release Period Results for Year 2000 format yyyy Lateral Distance from Source 280 m Depth into Low k Zone 3 m Return to Main Screen Figure 1 9 DSM Input Parameters Plume Zone Evaluation Calibrated MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 129 CASE STUDY 1B INDUSTRIAL SITE CONNECTICUT DANDY SALE MODEL Low k Zone Aqueous Concentration mg L for Year 2000 Lateral Distance from Source m 112 00 126 00 140 00 154 00 168 00 182 00 196 00 210 00 224 00 238 00 252 00 266 00 280 00 E o E x f o i o E a o a m0 19 m19 37 m37 56 m56 74 m74 93 m93 111 m111 130 m130 148 m148 167 m167 185 m185 204 9 See Trans Zone Aqueous Conc Export Low k 2 D Data HELP See Low k Aq Conc vs Dist M See Trans Zone Mass Discharge Aqueous Mass E Next Step See Low k Aq Conc vs Depth Return to DSM Data Input Sorbed Mass Save Data g t kg See Trans Zone Sorbed Conc See 2 D Low k Sorbed Conc Total Mass ad ka See Trans Zone Total Conc Return to Main Screen See 2 D Low k Total Conc Figure 1 10 DSM Output Plume Area Low k Zone Concentrations Calibrated MATRIX DIFFUSION TOOLKIT v USER S MANUAL Y 130 CASE
164. the Toolkit s SRM and DSM was obtained from Chapman and Parker 2005 MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 78 APPENDICES MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 79 APPENDIX A 1 SQUARE ROOT MODEL Appendix A 1 1 Estimation of Mass Discharge Purpose Determine the mass discharge output of the Square Root Model of the Matrix Diffusion Toolkit Note This derivation was originally developed by Parker et a 1994 and is detailed in AFCEE 2007 Given There is a finite amount of soluble organic constituents in the source zone in the dissolved sorbed and NAPL phases Assumptions The Toolkit uses a simplified conceptual model of a two layer aquifer system a transmissive layer and a low k layer where there are two different time periods 1 A loading period where there is a constant concentration of contaminants in the transmissive zone that drives contaminants into the low k zone 2 A release period where the transmissive zone is assumed to have no concentration and an upper range estimate of release from the low k zone is generated 3 The low k zone is at least 1 meter thick 4 There is no degradation in the low k zone Summary At any time t transverse diffusion of contaminants into the low k layer underlying the pool can be estimated using the equation ja RD RD vz Where Ma Mass discharge M T t Time since source was introduced T t Time at which source was removed from th
165. the source is active the low k aqueous concentration can be calculated at any time t using Sale et al 2008b b t 1 rpxHh xz t b 1 b b x do Jx h e erfc Gei 1 with I x z t b o Ds D Vo R R and y defined as Cx zt Co lt pum K Zz DESCH im C I x z t erfc Lx e gt 2 Me x y2 o Rr Ee SE SE exp du x 8 voie o E MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 85 APPENDIX A 2 DANDY SALE MODEL D Va De 5 D n D 6 y 9 where C x z t Aqueous concentration at lateral distance x depth z and time t M L C Mean plume loading concentration above the low k layer during the charging period M L Coefficient of transverse hydrodynamic dispersion L Source characteristic 1 L Effective transverse diffusion coefficient in the low k layer L T Effective molecular diffusion coefficient in the transmissive layer L T Molecular diffusion coefficient in free water L T Effective transverse diffusion coefficient in the transmissive layer L T Fraction organic carbon of the transmissive layer unitless Fraction organic carbon of the low k layer unitless Organic carbon partitioning coefficient L M Source zone length L Porosity of transmissive layer unitless Porosity of low k layer unitless Retardation factors for the transmissive layer unitless Retardation factor
166. timate the tortuosity of the low k materials where matrix diffusion has occurred diffusion coefficients and fraction organic carbon of the clays and silts being modeled etc The Toolkit provides default values and advice on selecting representative values for your site conditions Can the Toolkit be used for fractured rock sites Yes but the application and interpretation will require additional interpretation and expertise The model basically assumes a single transmissive zone which would be a fracture and a single low k zone the rock matrix To apply this to a fractured system the mass discharge and concentration would have to be multiplied by two to account for the contribution from both sides of the fracture To simulate multiple fractures you would have to multiply the results from a single fracture by the number of fractures contributing to the mass flux mass discharge at the point of interest What contaminants can be modeled with the Toolkit To date most of the research involving matrix diffusion processes for low k zones has focused on chlorinated solvents such as TCE trichloroethene and Methyl tert butyl ether MTBE However in theory matrix diffusion processes should apply to almost any dissolved contaminant including benzene and other aromatic compounds found in gasoline although the overall impacts may differ Matrix diffusion of dissolved metals and radionuclides can also be modeled if a simplifying assumption of a
167. to Enter Data Enter directly PARAMETER SOIL BULK DENSITY OF LOW k ZONE p Description Density of the saturated low k zone referred to as soil excluding soil moisture Typical Values Although this value can be measured in the lab estimated values are used in most cases A value of 1 7 g mL is used frequently for unconsolidated media Representative values for specific geologic media are shown below Lovanh et al 2000 derived from Domenico and Schwartz 1990 Clay 1 0 2 4 Loess 0 75 1 6 Sandstone 1 6 2 68 Shale 1 54 3 17 Limestone 1 74 2 79 Granite 2 24 2 46 Basalt 2 2 7 Medium Sand 1 34 1 81 MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 51 DSM DATA ENTRY Koerner 1984 reports these values for unit weight for saturated soils note no dry bulk density values are reported for these materials Glacial till very mixed grain 2 32 Soft glacial clay 1 77 Stiff glacial clay 2 07 Soft slightly organic clay 1 58 Soft very organic clay 1 43 Soft bentonite 1 27 Source of Data Either from an analysis of soil samples at a geotechnical lab or more commonly application of estimated values such as 1 7 g mL How to Enter Data Enter directly PARAMETER TRANSMISSIVE ZONE FRACTION ORGANIC CARBON foc Description Fraction of the aquifer material comprised of natural organic carbon in uncontaminated areas More natural organic carbon means higher adsorption of organic constituents on the aquifer matrix
168. to be entered directly into the white blue cells NOTE Although literature values are provided site specific hydrogeologic transport and plume characteristic values will likely provide better results If literature values are used and there is uncertainty in the value chosen sensitivity analyses should be conducted to determine the effects of the uncertainty on model predictions Recommendations regarding calibrating fitting the SRM to actual field data After the model has been set up and run model output can be compared to actual field data from monitoring wells using either a concentrations comparison or a mass discharge comparison Most times the initial run will not produce modeled data that match field data Considerations and recommended steps to improve the fit of simulated to field data are provided below The first caveat associated with calibrating the SRM is that the model assumes the original source zone is completely cleaned up and does not account for any residual source In other words at many sites the concentrations from matrix diffusion may only be causing part of the contaminant concentrations in monitoring wells residual mass from the source zone may also be contributing to the observed concentrations Consequently an exact match to observed concentration in a monitoring well should not be attempted if there is any uncertainty in matrix diffusion processes being the sole source of contaminants in the modeled zone
169. transmissive zone What s up with the gap Lower Range E MostLikely Upper Range B Field Data 1 00E 05 1 00E 04 EE D a t T ME M T 1 00E 02 EE ES 1 00E 01 4 1 00E 00 1 00E 01 1 00E 02 oo sz No or EE GE Ki E Q o ESD So E S3 Sc ke U EE o d ZS a 1 00E 03 1997 Re Plot Graph from Year 1996 toYear 2015 Next Step See Mass Discharge Results See Mass Results format yyyy format yyyy arat Soe DEES E Figure 1 3 SRM Output concentrations in the transmissive zone at location MW 01 The middle line is the most likely result from the SRM The square symbols with crosses are actual site data As can be seen this results in a very close match to actual field data MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 120 CASE STUDY 1B INDUSTRIAL SITE CONNECTICUT DANDY SALE MODEL B Dandy Sale Model DSM Input Data Data Type Source of Data Hydrogeology Trans zone description sand Boring logs Trans zone porosity 0 35 Site estimate Low k zone description silt Boring logs Low k zone porosity 0 43 Site estimate Trans zone seepage 0 37 m d Calculated based on site velocity estimates Transport Key constituent TCE Site history Low k Zone Mean concentration 1100 mg L Literature TCE solubility Molecular diffusion coefficient 9 1E 10 m sec Literature Toolkit default in fre
170. transverse dispersion and diffusion 7 There is no longitudinal dispersion in the transmissive layer 8 1 D transverse diffusion transport occurs in the low k layer 9 Retardation of contaminants in both layers is based on instantaneous equilibrium between aqueous and sorbed phases Summary Active Source Using a linear soil water partitioning coefficient the sorbed concentration in the low k layer at any time f can be calculated as C cornea x Z t C x Z5 t Ka 1 where C sorbea X Z t Sorbed concentration at lateral distance x depth z and time t M M C x Zt Aqueous concentration at lateral distance x depth z and time t M L calculated using Appendix A 2 1 Equation 1 Kg Soil water partitioning coefficient L M Tac Koc foc Fraction organic carbon of the low k layer unitless and Koc Organic carbon partitioning coefficient L M MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 88 APPENDIX A 2 DANDY SALE MODEL Exhausted Source Once the source is exhausted the low k sorbed concentration can be calculated at any time tas C sorbed x z t T z CG Z L t Ka 2 where C sorea X Z t t Sorbed concentration at lateral distance x depth Z and time t after the source has depleted M M and C x Z t t Aqueous concentration at lateral distance x depth z and time t after the source has depleted M L calculated using Appendix A 2 1 Equation 11 MATRIX DIFFUSION T
171. tribution function is 1 x A lt x lt B f x X where A and B are the lower and upper bounds respectively MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 112 CASE STUDIES MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 113 CASE STUDY 1A INDUSTRIAL SITE CONNECTICUT SQUARE ROOT MODEL CASE STUDY 1 INDUSTRIAL SITE CONNECTICUT Overview The Matrix Diffusion Toolkit was used to estimate the effects of diffusion into and from low k zones for the trichloroethene TCE plume at an industrial site in Connecticut Figure 1 1 Chapman and Parker 2005 have described the site in detail For this analysis both the Square Root Model SRM Case Study 1A and the Dandy Sale Model DSM Case Study 1B were applied as follows e Step 1 Initial values of all parameters obtained from either Chapman and Parker 2005 or default Matrix Diffusion Toolkit parameters were entered into the Toolkit e Step 2 Toolkit outputs were compared to field observed TCE concentrations This step was critical in determining how well default Toolkit parameters predicted actual field conditions e Step 3 Input parameters were adjusted as needed to improve the comparison with field observed TCE concentrations MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 114 CASE STUDY 1A INDUSTRIAL SITE CONNECTICUT SQUARE ROOT MODEL LEGEND Multilevel Bundle Conventional Well Plume N Groundwater Flow Building A Sheet Pile
172. ulated using Appendix A 2 7 n Porosity of transmissive layer unitless Pp Bulk density of transmissive layer M L Ky Soil water partitioning coefficient L M foc Koc MATRIX DIFFUSION TOOLKIT v USER S MANUAL Y 106 APPENDIX A 2 DANDY SALE MODEL foc Fraction organic carbon of the transmissive layer unitless and Koc Organic carbon partitioning coefficient L M MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 107 APPENDIX A 2 DANDY SALE MODEL Appendix A 2 10 Transmissive Layer Aqueous Mass Purpose Determine the transmissive layer aqueous phase mass output in the Dandy Sale Model of the Matrix Diffusion Toolkit Given The vertical plane source produces a plume in the transmissive zone that loads up the low k zone due to diffusion This vertical plane source is shut off and diffusion results in a release of contaminants from the low k zone Assumptions The model makes the following assumptions 1 A source considered to be a thin pool is introduced at the contact between the two layers upgradient of x 0 2 A loading period occurs where there is a constant concentration of contaminants in the transmissive zone that drives contaminants into the low k zone 3 A release period occurs where the transmissive zone is assumed to have no concentration and an upper range estimate of release from the low k zone is generated 4 There is no degradation in either layer 5 Both layers are uniform homo
173. ulic gradient data parameters effective porosity can be estimated How to Enter Data 1 Select units and enter directly or 2 Calculate by pressing the Calculate V button and entering values for a Hydraulic conductivity and b Hydraulic gradient PARAMETER TRANSMISSIVE ZONE HYDRAULIC CONDUCTIVITY K cm sec ft or m day ft or m yr Description Measure of the permeability of the transmissive layer To characterize concentrations in the transmissive layer representative measurements are required for the Darcy velocity or both the hydraulic flow gradient and the hydraulic conductivity of the flow system Representative measurements of the hydraulic conductivity of the transmissive layer should be obtained at one or more locations using appropriate slug test or pumping test methods Newell et al 2003 Typical Values Silts 1x10 4x1073 cm s Silty sands 1x10 9 1x1071 cm s Clean sands 1x10 3 1 cm s Gravels gt 1cm s Newell et al 1996 Source of Data Pump tests or slug tests at the site It is strongly recommended that actual site data be used for all matrix diffusion evaluations How to Enter Data 1 Select units and 2 Enter directly PARAMETER TRANSMISSIVE ZONE HYDRAULIC GRADIENT i Description The slope of the potentiometric surface In unconfined aquifers this is equivalent to the slope of the water table Typical Values 0 0001 0 1 ft ft 0 0001 0 1 m m Source of Data Calculated by construct
174. urn to Main Screen Approximate Width Width of Blue Box 3 00E 02 Figure 2 2 SRM Input Parameters Fluorescein Initial MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 137 CASE STUDY 2A SAND TANK STUDY SQUARE ROOT MODEL SRM Data Input Screen DATA INPUT INSTRUCTIONS Matrix Diffusion Toolkit Version 1 0 LLL Enter value directly Wl value calculated by Toolkit Do not enter data Site Location and ID Sand Tank stud 1 SYSTEM UNITS 2 ANALYSIS TYPE 5 PLUME CHARACTERISTICS CONT D SI Units O English Units C Source Zone Analysis Plume Analysis C PRB Analysis B Concentration of Contour Line in Blue Box 1 88E 00 Representative Concentration OK to Override 1 88E 00 Restore Uncertainty in Plume Concentration Estimations 3 HYDROGEOLOGY l Low k Zone Description B Low k Zone Porosity Transmissive Zone Darcy Velocity me Calculate va 4 TRANSPORT LOW k ZONE 6 GENERAL Key Constituent User Input z Source Loading Starts in Year 2006 format yyyy Molecular Diffusion Coefficient in Free Water Source Removed in Year format yyyy Apparent Tortuosity Factor Exponent H Retardation Factor CalculateR 5 PLUME CHARACTERISTICS See Back Diffusion Results from Year 2006 format yyyy to Year 2006 36 format yyyy in Intervals of 0 00274 yrs 7 FIELD DATA FOR COMPARISON High Concentration Zone Black Box in Picture Year Approximate Length Length of Black
175. versity of Guelph high resolution core sampling techniques Mr Mike Singletary of the Naval Facilities Engineering Command was the point of contact for this project e Due to availability of limited site information Toolkit default values were used as input parameters where necessary Seepage velocity initial source concentration and low k formation fraction organic carbon were varied until a reasonable comparison between simulated and observed concentrations was obtained at the three field observation locations e Toolkit inputs are shown on Figures 3 2 and 3 3 for the initial and calibrated models respectively A DSM output is shown on Figure 3 4 while comparisons of the Toolkit simulated with observed values are shown on Figures 3 4 3 5 and 3 6 e Torun the model hydrogeological data were entered in Section 2 transport parameters in Section 3 source zone characteristics in Section 4 and desired output information in Section 5 KEY POINT The Toolkit was able to reproduce observed soil concentrations to within an order of magnitude The initial site estimated seepage velocity of 20 ft yr was unable to reproduce the observed plume length at the site A better comparison between simulated and observed soil concentrations was obtained by increasing the seepage velocity initial source groundwater concentration low k zone fraction organic carbon and the diffusion coefficient Based on the calibrated model the Toolkit yielded a go
176. volume of the aquifer matrix Differences between total and effective porosity reflect lithologic controls on pore structure In unconsolidated sediments coarser than silt size effective porosity can be less than total porosity by 2 5 e g 0 28 vs 0 30 Smith and Wheatcraft 1993 For this implementation of the model effective porosity is typically used and is assumed to be similar to total porosity for mass and mass transfer calculations In other words to simplify the model both effective and total porosity are not entered separately but assumed to be the same value Typical Values Values for effective porosity Silt 0 01 0 30 Gravel 0 10 0 35 Fine Sand 0 10 0 30 Medium Sand 0 15 0 30 Coarse Sand 0 20 0 35 From Wiedemeier et al 1999 originally from Domenico and Schwartz 1990 and Walton 1988 Source of Data Typically estimated Occasionally obtained through physical property testing of site soil samples One commonly used value for silts and sands is an effective porosity of 0 25 The ASTM RBCA Standard ASTM 1995 includes a default value of 0 38 to be used primarily for unconsolidated deposits A collection of default values is presented in the Geologic Parameter Database included in this manual How to Enter Data Enter directly Note that if the transmissive zone description is selected from the drop down list the Toolkit provides a default value for the parameter MATRIX DIFFUSION TOOLKIT Y USE
177. w k layer at time t M M t Aqueous phase mass in the low k layer at time t M calculated using Appendix A 2 4 Equation 1 U n Porosity of low k layer unitless P Bulk density of low k layer M L Ka Soil water partitioning coefficient L M f oc Koc Foc Fraction organic carbon of the low k layer unitless and MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 94 APPENDIX A 2 DANDY SALE MODEL Koc Organic carbon partitioning coefficient L M Exhausted Source Once the source is exhausted the low k sorbed phase mass can be calculated at any time tas Mi T Mat dl 2 where M t Sorbed phase mass at time t after the source has depleted M and M 44 t z Aqueous phase mass at time t after the source has depleted M calculated using Appendix A 2 4 Equation 2 MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 95 APPENDIX A 3 DANDY SALE MODEL Appendix A 2 6 Low k Total Mass Purpose Determine the low k total mass output in the Dandy Sale Model of the Matrix Diffusion Toolkit Given The vertical plane source produces a plume in the transmissive zone that loads up the low k zone due to diffusion This vertical plane source is shut off and diffusion results in a release of contaminants from the low k zone Assumptions The model makes the following assumptions 1 A source considered to be a thin pool is introduced at the contact between the two layers upgradient of x 0 2 A load
178. were unavailable the maximum observed concentration in the plume zone as suggested by the Toolkit was used as the starting point for source concentrations e An uncertainty of a factor of 10 was assumed for concentration estimations e Asheet pile enclosure was installed in 1994 around the DNAPL area Figure 1 1 However to account for the travel of contaminated groundwater present at the time of the sheet pile an effective source removal time of 1996 was used in the Toolkit e Monitoring data from well MW 01 was used for calibration KEY POINTS The SRM of the Toolkit was able to reproduce observed groundwater concentrations to within an order of magnitude Use of site specific values documented by Chapman and Parker 2005 and Toolkit default values for molecular diffusion coefficient in free water and apparent tortuosity factor exponent provided a reasonable comparison to actual observed TCE concentrations in MW 01 Figure 1 3 Therefore no adjustment of any input parameters was necessary Chapman and Parker 2005 estimated a total mass in the aquitard of 3000 kg for the year 2000 Comparably the Toolkit estimates a most likely mass of 1414 kg This is well within the order of magnitude level of accuracy goal for this model Note that a typical advection dispersion type model would show no mass in the low k unit a fundamentally incorrect conceptual model We feel using a simple model to get MATRIX DIFFU
179. x __Calculatev al TRANSPORT Key Constituent enter directly or choose from drop down list TCE Plume Loading Concentration Immediately Above Low k Zone in Vertical Plane Source During Loading Period 1100 mo v Molecular Diffusion Coefficient in Free Water 9 10E 10 m2 sec B Transmissive Zone Apparent Tortuosity Factor Exponent 9 Low k Zone Apparent Tortuosity Factor Exponent 33 4 SOURCE ZONE CHARACTERISTICS Bulk Density of Transmissive Zone g mL Source Zone Length 32 1 m Bulk Density of Low k Zone g mL Source Zone Width 39 3 m Transverse Vertical Hydrodynamic 1 00E 03 m Restore H Distribution Coefficient mL g Source Loading Starts in Year 1952 format yyyy or Calculated R Source Removed in Year 1997 format yyyy Transmissive Zone Fraction of Organic Carbon 3 80E 04 1 17 5 GENERAL Low k Zone Fraction of Organic Carbon 5 40E 04 1 18 See Release Period Results for Organic Carbon Partitioning Coefficient 9 33E 01 L kg Year 1997 format yyyy Lateral Distance from Source 0 001 m Depth into Low k Zone 3 m VEU Paste Example Se Show Graph Return to Model Selection Screen Return to Main Screen Figure 1 4 DSM Input Parameters Source Zone Evaluation MATRIX DIFFUSION TOOLKIT v USER S MANUAL Y 124 CASE STUDY 1B INDUSTRIAL SITE CONNECTICUT DANDY SALE MODEL Low k Zone Aqueous Concentration mg L at x 0 001 m for Y
180. x and time t Cwen at 7 N 3 Well concentration at time t after the source is exhausted is obtained by subtracting the well concentration calculated in Step 2 from Step 1 Numerical Integration Method The Matrix Diffusion Toolkit employs a 10 pt Gaussian quadrature to solve polynomials MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 104 APPENDIX A 2 DANDY SALE MODEL Appendix A 2 8 Transmissive Layer Sorbed Concentration Purpose Determine the transmissive layer sorbed concentration output in the Dandy Sale Model of the Matrix Diffusion Toolkit Given The vertical plane source produces a plume in the transmissive zone that loads up the low k zone due to diffusion This vertical plane source is shut off and diffusion results in a release of contaminants from the low k zone Assumptions The model makes the following assumptions 1 A source considered to be a thin pool is introduced at the contact between the two layers upgradient of x 0 2 A loading period occurs where there is a constant concentration of contaminants in the transmissive zone that drives contaminants into the low k zone 3 A release period occurs where the transmissive zone is assumed to have no concentration and an upper range estimate of release from the low k zone is generated 4 There is no degradation in either layer 5 Both layers are uniform homogeneous isotropic and infinite in the z direction perpendicular to groundwater flow 6 One d
181. y function is x o gt 0 o o f ee YO 27T 20 where o is the standard deviation and u the mean of the underlying normal distribution Lognormal distributions are typically specified in two ways throughout literature Swiler and Wyss 2004 One way as described above is to use the mean and standard deviation of the underlying normal distribution The other way is to use the mean of the lognormal distribution a and a term called the Error Factor For a lognormal distribution the error factor is the ratio of the 95 percentile to the median or equivalently the ratio of the median to the 5 percentile Therefore the error factor represents the width of a 9096 confidence interval around the median In terms of the error factor the relationship between the underlying normal distribution and the lognormal distribution can be described by o In error factor 1 645 and MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 111 APPENDIX A 3 PROBABILITY DISTRIBUTIONS 2 Oo In g 4 a 5 where o is the mean of the lognormal distribution and o and u the standard deviation and mean of the underlying normal distribution respectively The Toolkit describes the lognormal distribution using the error factor A 3 3 Uniform Distributions A uniform distribution is specified over a particular interval and implies that all the points within that interval have equal probability of occurring The uniform probability dis
182. year that best represents when concentrations in the middle of the modeled area were reduced significantly by source remediation or 2 when source zone natural attenuation processes reduced the concentrations in the middle of the modeled area significantly For example the source could likely be considered removed by natural attenuation for the purposes of this model if the transmissive zone of the modeled area has been reduced by 90 or 99 compared to the historical all time concentrations This can be used as a calibration parameter see the beginning of this section How to Enter Data Enter directly PARAMETER SEE RELEASE PERIOD RESULTS FROM YEAR Starting year for displaying matrix diffusion results How to Enter Data Enter directly Step 7 Field Data for Comparison PARAMETER FIELD DATA FOR COMPARISON YEAR Description Years in which field data are available for calibration These data are displayed with model results in the Next Step Show Graph option Source of Data Monitoring wells located in the area of interest MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 35 SRM DATA ENTRY How to Enter Data Enter directly PARAMETER FIELD DATA FOR COMPARISON CONCENTRATION Description Concentration measurements in transmissive zone area of interest These data are displayed with model results in the Next Step Show Graph option Typical Values 0 001 10 000 mg L Source of Data Monitoring wells located in the a
183. zone length L n Porosity of transmissive layer unitless n Porosity of low k layer unitless R Retardation factors for the transmissive layer unitless R Retardation factors for the low k layer unitless t Time since source was introduced T V Seepage velocity in the transmissive layer L T Vo Contaminant transport velocity in the transmissive layer L T x Lateral distance from source edge L and Dummy integration variable Positive mass flux values indicate diffusion from the transmissive zone into the low k zone Negative values indicate diffusion from the low k zone into the transmissive zone While the source is on diffusion will occur from the transmissive zone into the low k zone Concentration in the transmissive layer can be estimated by imagining the contaminant plume as a train car picking up initial mass from the source and losing mass to the low k layer through diffusion as shown on Figure A 2 7 2 Initial loading from Figure A 2 7 2 Schematic for calculating concentration in the transmissive zone Concentrations are calculated using the following steps 1 Assume a monitoring well with a 10 ft 3 m screened interval located at the distance x of interest MATRIX DIFFUSION TOOLKIT Y USER S MANUAL Y 101 APPENDIX A 2 DANDY SALE MODEL 2 Set well concentration to zero if point of interest is greater than the point of plume arrival e x 2 Vt where V is the contaminant v

Matrix Diffusion Toolkit User`s Manual - CLU-IN

Contents

Download Pdf Manuals

Related Search

Related Contents