UNITED STATES IHV1RONMENTAL PROTECTION AGENCY
                              WASHINGTON B.C. 20460
                                  August 18, 1992
                                                                      OFFICE OF
                                                                   THE ADMINISTRATOR
                                                                 SCIENCE ADVBOKY BOARD
EPA-SAB-DWC-LTR-92-013

Honorable William K, Reilly
Administrator
U.S. Environmental Protection Agency
401 M Street, SW
Washington, DC 20460

             Subject:      Science Advisory Board Review of the Viral Transport
                          (VIRALT) Model

Dear Mr. Reilly:

       On December  6-7, 1990, the Drinking Water Committee (DWC) of the 1PA Science
Advisory Board (SAB) reviewed VIRALT, a modular semi-analytical and numerical model
for simulating transport and fate of viruses in ground water. The stated purpose for
developing the model was for use as a screening tool for assisting in the development and
implementation of the forthcoming ground water disinfection rule.  The Committee
provided oral comment to the Agency at that public meeting.  The comments which form
the basis for this report were drafted following the meeting. At a subsequent meeting of
the Committee on February 11-12,  1992, Agency staff requested that the Committee's
advice concerning the model be formalized in a Science Advisory Board report.

       The most serious deficiency of the model development is the lack of field validation.
Until this field validation is completed, the model should not be used.  Default values for
source concentrations, adsorption coefficients and other parameters  have been taken from
the literature and are  not well supported or documented.  Furthermore, the assumptions
used to develop the model must also be validated.  Without these validations, VIRALT
cannot be used to identify wells at risk or the disinfection levels required. A very
important issue that needs to be addressed is definition of the target user group for
VIRALT.  The documentation states that it is intended to be used by EPA, State and local
technical staff.  One potential problem that this presents is the wide range of technical

-------
capabilities that would be represented within such a diverse group.  The documentation is
not adequate for users with a cursory knowledge of groundwater flow and transport
phenomena.

      The Committee was asked to: 1) make a thorough review of the appropriateness of
the modeling assumptions and suggest possible revisions that might improve the predictive
capabilities of the model, and discuss the limitations of the model, and 2) review the
validity and utility of the model.

                        1. SOENTIEie AREAS OF REVIEW

1.1   Assumptions Underlying the VIRAL.TModel an4 LJn*it|iiQtis fhereof

      VIMALT is based on two sets of assumptions: 1) those governing groundwater flow
computations, and 2)  those governing viral transport computations.  Specific EPA questions
concerning these assumptions are;

      a)  "  Are the assumptions made in the development of the code appropriate
             considering the intended use of the model and the limitations of the available
             data?

      b)     Are any of the assumptions oversimplified or unacceptable? If so, what steps
             should  be taken to address concerns about assumptions?

      1.1.1  Groundwater How Assumptions

      VIRALT employs a semi-analytical flow module for performing groundwater flow
computations.

      a)     Two Dimensional glow - The flow module is based on two dimensional flow
             in a confined (or shallow unconfined) aquifer with fully penetrating wells,
             streams, and boundaries.  This should generally produce acceptable
             predictions for shallow aquifers that would be most likely to be contaminated
             by septic tanks or leaky sewers. The vadose zone is ignored so the model
             would tend to produce overly conservative predictions for deep aquifers with
             thick unsaturated zones because flow in the vertical direction would be

-------
       significant and would result in longer travel times than predicted by the
       model.

       Partially penetrating wells or stream boundaries in relatively deep aquifers
       would result in significant vertical flow that, in turn, would result in higher
       flow velocities than the model would predict leading to non-conservative
       model predictions.  Related issues that need to be clarified art: 1) how the
       model teats injection of virus into the aquifer, 2) are they mixed over the
       depth of the aquifer or are they assumed to be injected at the ground water
                                 * • •
       surface as would most likely be the case in a real situation, and 3) what
       initial virus concentrations  should be assumed?

b)     Steady State, Flow Field - The steady state assumption is probably reasonable
       for a screening model with the limited data that are available for most sites,
       Modifying the model to simulate transient flow conditions would require a
       substantial increase in computational effort and data requirements and would
       not be justified in most cases. The limitations of the steady state assumption
       should be discussed in more detail in the documentation to aid the user in
       understanding its implications for a particular site.

c)     Homogenous Isotropic Aquifer - This assumption would be reasonable for
       situations where application of the model would be most appropriate,  i.e.,
       shallow fine grained aquifers. It is not appropriate for layered, or for Karst
       and fractured rock aquifers where short eircuitinf or channeling might occur.
       In these cases it would  result in non-conservative predictions of flow times.

1.1.2  Viral Transport Assumptions

a)     Viral Transport Mechanism - The transport equation is based on the
       assumption that viruses behave like dissolved solutes, but viruses are known
       to behave Ike charged colloids.

b)     Multiple Default Parameters - The user is presented with multiple sorption
       and inactivation coefficients for 5 viruses with no guidance on parameter
       selection.  Consideration should be  given to selecting a standard or indicator
       virus, i.e. MS-2 phage.

-------
c)     Transverse Dispersion Neglected - The transport module is based on the
       assumption that viral transport occurs along groundwater pathlines.
       Dispersion perpendicular to the direction of travel is ignored and is
       represented in the documentation as being conservative because viral
       concentrations along pathlines would not be diluted by the effects of
       transverse dispersion.  In reality, adjacent wells might be subject to viral
       contamination because of dispersion of the viral plume, although this is not
       predicted by the model.

d)     Viral Adsorption - The model uses a linear adsorption isotherm that the
       available data do not seem to support.  A more thorough study of available
       data should be undertaken to assess the validity of the linear isotherm
       approach.  Non-linear or non-equilibrium sorption processes might yield
       higher virus concentrations at extraction wells than predicted using the linear
       isotherm.  The model does not consider desorptkm that might in some cases
       to be important.  VIRALT could be probably used to identify potential source
       peas, but EPA already has models with that capability (GWPATH or EESSQ
       with modification).

e)     First Order Viral Inactivation - First order viral motivation is not adequately
       justified, and the basis for temperature correction is not given. The model
       assumes that viral inactivation rates are the same for viruses in pore water
       and adsorbed on soil particles.  Viruses are inactivated at different rates in
       adsorbed and dispersed states.

Validation and Utility of Viralt

       Model Validation - The most serious deficiency is the lack of field
       validation and the model should not be used until this field validation
       is completed. Default values for source concentrations, adsorption
       coefficients and other parameters have been taken from the literature
       and are not well supported or documented.  Furthermore, the
       assumptions used to develop the model must also be validated.
       Without these validations, VIKALT cannot be used to identify wells at
       risk or the disinfection levels required.

-------
L3    Other Issues

       A very important issue that need to be addressed is definition of the target user
group for VTRALT,  The documentation states that it is intended to be used by EPA, State
and local technical staff.  One potential problem that this presents is the wide range of
technical capabilities that would be represented by such a diverse group.  The
documentation is not adequate for users with a cursory knowledge of groundwater flow and
transport phenomena.

       Another important issue is the multiplicity of choices of viruses, soil types, aquifer
characteristics, etc., with which the user is faced.   U.S. EPA should  consider adoption of
a standard or indicator virus and collecting data for this virus over a wide range of
conditions.

                                 2,  CONCLUSIONS

       The following general conclusions can be made regarding the model.

2,1    Model Assumption^ find Limitations

       a)     Steady State Assumption - Limitations of the steady state assumption need to
             be described  in more detail to aid the user in applying the model to a
             particular site.

       b)     Fully Penetrating Weils/Boundaries - The differences of fully versus partially
             penetrating wells need to be described in detail for the benefit of the typical
             user. Another issue that needs clarification for all users is that of source
             injection. Is the contaminant source introduced at the groundwater surface or
             uniformly over the depth of the aquifer and what concentration of virus are
             introduced?

       c)     Aquifer and Virus Transport Parameters - The user's manual should provide
             the typical user more guidance on selection of modeling parameters.  Virus
             transport parameters are tabulated in Appendix A, but the user  is given little
             guidance in selecting appropriate values for specific site conditions.

-------
       d)     Interpretation of Results - The typical user is left to Ms or her own devices in
             interpreting model results.  No mention is made of results interpretation in
             the users manual,

       e)     Model Applicability - The model is limited to predicting viral transport in
             shallow  homogenous aquifers.

       f)     Vadose  Zone - VIRALT will produce overly conservative results in deep
             aquifers with thick un-saturated zones. Long flow times and high rates of
             viral sorption/inactivation will produce viral concentrations in the aquifer
             significantly below model predictions.

       g)     Homogeneous Aquifers - VIRALT is suited neither for karst nor fractured
             rock aquifers and  will generally make non-conservative predictions of virus
             removal in those formations.

       h)     Linear Viral Adsorption Isotherm - The linear isotherm is not justified based
             on the available data,

2.2    Validation and  Utility of  VIRALT

       Before a model can be considered reliable and useful, it must be validated for the
conditions that it will be expected to simulate.

       a)     Validation -  VMALT has not been field validated to date.  A field validation
             must be performed for aquifer systems conforming as closely as possible to
             the assumptions under which the model was developed.

       b)     Identification of High Risk Wells - VIRALT can probably be used  to identify
             high risk wells.

       c)     goprce Identification Capabilities - VIRALT cannot be used to reliably
             identify  contaminant sources.  EPA already has models with that capability
             (GWPATH or 1ESSQ with modification).

-------
2-3   Documentation

      The User's Manual is generally adequate for the experienced modeler who Is fully
aware of the limitations of the model and is experienced in selecting appropriate model
coefficients and parameters.  To make VTRALT useful for the inexperienced user, however,
deficiencies in the User's Manual need to be resolved.

                            3, SECOMMENDATIONS

      The following recommendations are made based on our review of VffiALT.

3.1   Validation of the Model

      The model must be validated in the field before adoption for routine use.  It has
never been validated by comparing its predictions with field data on virus transport in
groundwater. As part of the validation, the appropriateness of the assumptions that have
been made and the values of the parameters assumed must be evaluated.  These include:

      a)     Steady-state assumption

      b)     Fully penetrating well assumption

      c)     Saturated zone assumption

      d)     Virus adsorption and inactivatbn parameters

      e)     Initial virus concentration and means of introduction of virus to the aquifer

      f)     Selection of target virus assumption

3-2   Define Target User Group

      The target user audience needs to be defined and the documentation revised
accordingly.   An experienced groundwater professional can use the model with the existing
documentation with the following recommended modifications. Before inexperienced users

-------
can productively use the model, however, documentation will require extensive revision and
the program made more user friendly,

       a)     Documentation - Deficiencies in the User's Manual need to be corrected.
             Some of the limitations of the model need to be described in more detail,

       b)     Interpretation of Results - The user should be given guidance in interpreting
             model results,

       We appreciate the opportunity to review this model and look forward to your written
response to the advice contained in this letter.
                                              Sincerely,
                                              Dr.
                                              Science Advisory Board
                                              Dr. Verne A. Ray, Cl
                                              DrinMng Water Committee
                                         8

-------
                   ENVIRONMENTAL PROTECTION AGENCY
                          SCIENCE ADVISORY BOARD
                        DRINKING WATER COMMITTEE
CHAIRMAN
      Dr. Vane A. Ray, Medical Research Laboratory, Pfizer Inc., Groton, Connecticut
VICE CHAIRMAN
      Dr. Vem L. Snoeyink, Department of Civil Engineering, University of Illinois, Urbana,
Illinois                                ,  .
MR^mERS/CONSTJLTANTS
      Dr. Richard J, Bull, College of Pharmacy,  Washington State University, Pullman,
Washington
      Dr. Gary P. Carlson, Department of Pharmacology and Toxicology, School of Pharmacy,
Purdue University, West Lafayette, Indiana
      Dr. Keith E. Cams, East Bay Municipal Utility District,  Oakland,  California
      Dr. Lenore S. Oesceri, Rensselaer Polytechnic Institute, Materials Research Center,
Troy, New York
      Dr. David G. Kaufman, Department of Pathology, University of North Carolina, Chapel
Hill, North Carolina
      Dr. Ramon G. Lee, American Water Works Service Company, Voorhees, New Jersey
      Dr. Edo D. Pellizzari, Research Triangle Institute, Research Triangle Park, North
Carolina
      Dr. Mark D. Sobsey, Department of Environmental Sciences and Engineering, School
of Public Health, University of North Carolina, Chapel HiE, North Carolina
      Dr. James M. Symons, Department of Civil and Environmental Engineering, University
of Houston, Houston, Texas
SCIENCE ADVISORY BOARD STAFF
      Mr. A. Robert Flaak, Assistant Staff Director and Acting Designated Federal Official,
Science Advisory Board (A-101F), U.S. EPA,  401 M Street, SW, Washington,  DC 20460
      Mrs. Frances Dolby, Staff Secretary, Drinking Water Committee, Science Advisory
Board (A-101F), U.S. EPA, 401 M Street, SW, Washington, DC 20460

-------
                                       NOTICE

    This report has been written as a part of the activities of the Science Advisory Board, a
public advisory group providing extramural scientific information and advice to the
Administrator and other officials of the Environmental Protection Agency.  The Board is
structured to provide balanced, expert assessment of scientific matters related to problems
facing the Agency.  This report has not been reviewed for approval by the Agency and,
hence, the contents of this report do not necessarily represent the views and policies of the
Environmental Protection Agency, nor of other agencies in the Executive Branch of the
Federal  government, nor does mention of trade "names or commercial products constitute a
recommendation for use.

-------
                                Distribution List
Administrator
Deputy Administrator
Assistant Administrators
Deputy Assistant Administrator for Research and DeYelopment
Deputy Assistant Administrator for Water
EPA Regional Administrators
EPA Laboratory Directors
EPA Headquarters Library                ,.
EPA Regional Libraries
EPA Laboratory Libraries

-------