[House Hearing, 108 Congress]
[From the U.S. Government Publishing Office]





   FOLLOWING TOXIC CLOUDS: SCIENCE AND ASSUMPTIONS IN PLUME MODELING

=======================================================================

                                HEARING

                               before the

                   SUBCOMMITTEE ON NATIONAL SECURITY,
                   EMERGING THREATS AND INTERNATIONAL
                               RELATIONS

                                 of the

                              COMMITTEE ON
                           GOVERNMENT REFORM

                        HOUSE OF REPRESENTATIVES

                      ONE HUNDRED EIGHTH CONGRESS

                             FIRST SESSION

                               __________

                              JUNE 2, 2003

                               __________

                           Serial No. 108-91

                               __________

       Printed for the use of the Committee on Government Reform


  Available via the World Wide Web: http://www.gpo.gov/congress/house
                      http://www.house.gov/reform


                                 ______

91-396              U.S. GOVERNMENT PRINTING OFFICE
                            WASHINGTON : 2003
____________________________________________________________________________
For Sale by the Superintendent of Documents, U.S. Government Printing Office
Internet: bookstore.gpo.gov  Phone: toll free (866) 512-1800; (202) 512ï¿½091800  
Fax: (202) 512ï¿½092250 Mail: Stop SSOP, Washington, DC 20402ï¿½090001

                     COMMITTEE ON GOVERNMENT REFORM

                     TOM DAVIS, Virginia, Chairman
DAN BURTON, Indiana                  HENRY A. WAXMAN, California
CHRISTOPHER SHAYS, Connecticut       TOM LANTOS, California
ILEANA ROS-LEHTINEN, Florida         MAJOR R. OWENS, New York
JOHN M. McHUGH, New York             EDOLPHUS TOWNS, New York
JOHN L. MICA, Florida                PAUL E. KANJORSKI, Pennsylvania
MARK E. SOUDER, Indiana              CAROLYN B. MALONEY, New York
STEVEN C. LaTOURETTE, Ohio           ELIJAH E. CUMMINGS, Maryland
DOUG OSE, California                 DENNIS J. KUCINICH, Ohio
RON LEWIS, Kentucky                  DANNY K. DAVIS, Illinois
JO ANN DAVIS, Virginia               JOHN F. TIERNEY, Massachusetts
TODD RUSSELL PLATTS, Pennsylvania    WM. LACY CLAY, Missouri
CHRIS CANNON, Utah                   DIANE E. WATSON, California
ADAM H. PUTNAM, Florida              STEPHEN F. LYNCH, Massachusetts
EDWARD L. SCHROCK, Virginia          CHRIS VAN HOLLEN, Maryland
JOHN J. DUNCAN, Jr., Tennessee       LINDA T. SANCHEZ, California
JOHN SULLIVAN, Oklahoma              C.A. ``DUTCH'' RUPPERSBERGER, 
NATHAN DEAL, Georgia                     Maryland
CANDICE S. MILLER, Michigan          ELEANOR HOLMES NORTON, District of 
TIM MURPHY, Pennsylvania                 Columbia
MICHAEL R. TURNER, Ohio              JIM COOPER, Tennessee
JOHN R. CARTER, Texas                CHRIS BELL, Texas
WILLIAM J. JANKLOW, South Dakota                 ------
MARSHA BLACKBURN, Tennessee          BERNARD SANDERS, Vermont 
                                         (Independent)

                       Peter Sirh, Staff Director
                 Melissa Wojciak, Deputy Staff Director
                      Rob Borden, Parliamentarian
                       Teresa Austin, Chief Clerk
              Philip M. Schiliro, Minority Staff Director

 Subcommittee on National Security, Emerging Threats and International 
                               Relations

                CHRISTOPHER SHAYS, Connecticut, Chairman

MICHAEL R. TURNER, Ohio
DAN BURTON, Indiana                  DENNIS J. KUCINICH, Ohio
STEVEN C. LaTOURETTE, Ohio           TOM LANTOS, California
RON LEWIS, Kentucky                  BERNARD SANDERS, Vermont
TODD RUSSELL PLATTS, Pennsylvania    STEPHEN F. LYNCH, Massachusetts
ADAM H. PUTNAM, Florida              CAROLYN B. MALONEY, New York
EDWARD L. SCHROCK, Virginia          LINDA T. SANCHEZ, California
JOHN J. DUNCAN, Jr., Tennessee       C.A. ``DUTCH'' RUPPERSBERGER, 
TIM MURPHY, Pennsylvania                 Maryland
WILLIAM J. JANKLOW, South Dakota     CHRIS BELL, Texas
                                     JOHN F. TIERNEY, Massachusetts

                               Ex Officio

TOM DAVIS, Virginia                  HENRY A. WAXMAN, California
            Lawrence J. Halloran, Staff Director and Counsel
              Kristine McElroy, Professional Staff Member
                        Robert A. Briggs, Clerk
                    David Rapallo, Minority Counsel


                            C O N T E N T S

                              ----------                              
                                                                   Page
Hearing held on June 2, 2003.....................................     1
Statement of:
    Rhodes, Keith, Chief Technologist, General Accounting Office; 
      Anna Johnson-Winegar, Deputy Assistant to the Secretary of 
      Defense for Chemical, Biological Defense Programs; Donald 
      L. Ermak, Program Leader, National Atmospheric Release 
      Advisory Center, Lawrence Livermore Laboratory; Bruce 
      Hicks, Director, Air Resources Laboratory, National Oceanic 
      and Atmospheric Administration; Eric Barron, Chair, Board 
      on Atmospheric Sciences and Climate, National Research 
      Council; and Steven R. Hanna, adjunct associate professor 
      of Harvard School of Public Health.........................     7
Letters, statements, etc., submitted for the record by:
    Barron, Eric, Chair, Board on Atmospheric Sciences and 
      Climate, National Research Council, prepared statement of..   105
    Byrd, Hon. Robert C., a Senator in Congress from the State of 
      West Virginia, prepared statement of.......................   120
    Ermak, Donald L., Program Leader, National Atmospheric 
      Release Advisory Center, Lawrence Livermore Laboratory, 
      prepared statement of......................................    82
    Hanna, Steven R., adjunct associate professor of Harvard 
      School of Public Health, prepared statement of.............   111
    Hicks, Bruce, Director, Air Resources Laboratory, National 
      Oceanic and Atmospheric Administration, prepared statement 
      of.........................................................    95
    Johnson-Winegar, Anna, Deputy Assistant to the Secretary of 
      Defense for Chemical, Biological Defense Programs, prepared 
      statement of...............................................    66
    Rhodes, Keith, Chief Technologist, General Accounting Office, 
      prepared statement of......................................    10
    Shays, Hon. Christopher, a Representative in Congress from 
      the State of Connecticut, prepared statement of............     3

 
   FOLLOWING TOXIC CLOUDS: SCIENCE AND ASSUMPTIONS IN PLUME MODELING

                              ----------                              


                          MONDAY, JUNE 2, 2003

                  House of Representatives,
Subcommittee on National Security, Emerging Threats 
                       and International Relations,
                            Committee on Government Reform,
                                                    Washington, DC.
    The subcommittee met, pursuant to notice, at 1 p.m., in 
room 2154, Rayburn House Office Building, Hon. Christopher 
Shays (chairman of the subcommittee) presiding.
    Present: Representatives Shays, and Turner.
    Staff present: Lawrence Halloran, staff director and 
counsel; Kristine McElroy, professional staff member; Robert A. 
Briggs, clerk; David Rapallo, minority counsel; and Jean Gosa, 
minority assistant clerk.
    Mr. Shays. A quorum being present, the Subcommittee on 
National Security, Emerging Threats and International 
Relations' hearing entitled, ``Following Toxic Clouds, Science 
and Assumptions in Plume Modeling,'' is called to order.
    What is the difference between an estimate and a guess? 
When plotting the path of a chemical, biological or 
radiological plume, the difference between a reasonable 
approximation and an unwarranted assumption can mean life or 
death.
    For U.S. troops on foreign battlefields, and for civilians 
here at home, the science of dispersion modeling lies at the 
heart of current efforts to prepare for, respond to, and 
recover from toxic attacks. From the trenches of World War I, 
through last months TOPOFF2 Exercise, military planners and 
homeland security officials have been attempting to refine the 
data and calculations needed to map the trajectory of noxious 
clouds.
    But, the variability of modeling techniques and the paucity 
of real-time data on weather patterns and weapon potency still 
makes projections too slow and limited to be relied upon for 
many critical decisions.
    Past attempts to model plume courses and concentrations 
yield important lessons and warnings. In 1996, this 
subcommittee heard persuasive testimony that coalition bombing 
of Iraqi chemical weapons facilities during the first Gulf war 
launched plumes that traversed large portions of the combat 
theater.
    Analysis of infrared satellite imagery and available 
weather data suggested broad dispersion patterns that would 
account for chemical agent detections at the time, detections 
once discounted but later deemed credible by the Department of 
Defense [DOD].
    But subsequent modeling of U.S. demolition of chemical 
weapons at Khamisiyah in Iraq conducted by DOD and the Central 
Intelligence Agency [CIA], between 1996 and 2000, produced 
varied yet uniformly narrower zones of risk than seemed 
plausible.
    So we asked the General Accounting Office [GAO], to review 
the Khamisiyah plume models and report on the implications of 
that process for Gulf war veterans and for all of those who 
might find themselves in the path of poisonous plumes at home 
or abroad in the future.
    The GAO findings highlight the dangers of reaching 
conclusions when critical data elements remain speculative or 
incomplete. According to GAO, DOD lacked essential information 
on the quantity and physical characteristics of the agents 
dispersed.
    Climate data was deficient. Arbitrary limits were placed on 
estimated plume altitudes, serious skewing downrange 
projections. DOD combined several in-house systems rather than 
select one validated modeling approach in the apparent hope 
that cumulative strengths would outweigh combined weaknesses. 
But, at some point, even that attempt, to err on the side of 
caution, produced more error than caution.
    Drawing cohorts based on flawed DOD modeling, 
epidemiological studies comparing exposed and unexposed 
veterans may be invalid.
    Once again, the benefit of any doubts about the extent of 
exposure risks has not gone to veterans, who now must bear the 
burden of proving themselves wrongly categorized by speculative 
Pentagon plume mapping.
    The same dangers and more confront dispersion modeling 
applications to meet homeland security requirements. Numerous 
special purpose models can produce very different outcomes 
using the same data.
    More vexing, very little data on wind and weather patterns 
has been captured in urban settings, the most inviting 
landscape for a terrorist attack. In the cold war, global and 
national security demanded the ability to plot the trajectory 
of ballistic missiles.
    In the war against weapons of mass destruction, we need to 
be able to predict the path of toxic clouds across new 
battlefields abroad and here at home.
    Today we examine efforts, past and present, to advance the 
science and perfect the art of plume modeling. Our panel of 
witnesses brings very impressive credentials and expertise to 
this discussion of a critical force projection and homeland 
security tool.
    We welcome them and we look forward to their testimony.
    [The prepared statement of Hon. Christopher Shays follows:]

    [GRAPHIC] [TIFF OMITTED] T1396.001
    
    [GRAPHIC] [TIFF OMITTED] T1396.002
    
    [GRAPHIC] [TIFF OMITTED] T1396.003
    
    Mr. Shays. At this time, the Chair would be happy to 
recognize Mr. Turner, the vice chairman of the subcommittee.
    Mr. Turner. Thank you, Mr. Chairman.
    I want to thank our witnesses and our chairman for having 
this important hearing.
    Plume modeling clearly has the potential for great 
usefulness in both issues of evacuation and first responders to 
terrorist attacks or industrial accidents. However, 
decisionmaking on current plume modeling may be premature.
    Another issue that I think needs to be addressed, I am 
looking forward to testimony today, on how plume modeling, once 
perfected, can be communicated to first responders through 
Federal, State and local governments so that it may be useful 
when an incident may be facing them. Thank you.
    Mr. Shays. Thank the gentleman. At this time, we will 
recognize our witnesses, and then swear them in and then begin 
the testimony.
    Our witnesses, beginning, and this is the order in which 
you will testify as well.
    Mr. Keith Rhodes, Chief Technologist, General Accounting 
Office; Dr. Anna Johnson-Winegar, Deputy Assistant to the 
Secretary of Defense for Chemical, Biological Defense Programs, 
Department of Defense; Dr. Donald L. Ermak, the program leader, 
National Atmospheric Release Advisory Center, Lawrence 
Livermore Laboratory. Mr. Bruce Hicks, Director, Air Resources 
Laboratory, National Oceanic and Atmospheric Administration; 
Dr. Eric Barron, Chair, Board on Atmospheric Sciences and 
Climate, National Research Council; and Dr. Steven R. Hanna, 
Adjunct Associate Professor of Harvard School of Public Health.
    If you would rise. And if there is anyone--is there anyone, 
Dr. Winegar, you more than others, that someone might testify. 
Or if so, if anyone else is accompanying you that might 
participate, I would prefer they stand up, even if they aren't 
ultimately called, so we don't have to swear them in twice.
    So if you would rise and raise your right hands please.
    [Witnesses sworn.]
    Mr. Shays. Note for the record, all of our witnesses have 
responded in the affirmative. I am going to ask Mr. Turner to 
take over. Mr. Rhodes, you will begin. May I just say, I am 
sorry, let me just make this point. We have 5 minutes. We have 
six panelists.
    We roll over. And you can take the other full 5 minutes, 
but we would prefer that the rollover, that you don't go too 
much further into that 5 minutes. I start to get a little 
nervous around 7 or 8 minutes. What we are finding is all of 
our witnesses are now spending 10 minutes. So that would be 
what I would hope would happen.

    STATEMENTS OF KEITH RHODES, CHIEF TECHNOLOGIST, GENERAL 
 ACCOUNTING OFFICE; ANNA JOHNSON-WINEGAR, DEPUTY ASSISTANT TO 
   THE SECRETARY OF DEFENSE FOR CHEMICAL, BIOLOGICAL DEFENSE 
PROGRAMS; DONALD L. ERMAK, PROGRAM LEADER, NATIONAL ATMOSPHERIC 
 RELEASE ADVISORY CENTER, LAWRENCE LIVERMORE LABORATORY; BRUCE 
HICKS, DIRECTOR, AIR RESOURCES LABORATORY, NATIONAL OCEANIC AND 
   ATMOSPHERIC ADMINISTRATION; ERIC BARRON, CHAIR, BOARD ON 
 ATMOSPHERIC SCIENCES AND CLIMATE, NATIONAL RESEARCH COUNCIL; 
  AND STEVEN R. HANNA, ADJUNCT ASSOCIATE PROFESSOR OF HARVARD 
                    SCHOOL OF PUBLIC HEALTH

    Mr. Rhodes. Yes, sir. Mr. Chairman, members of the 
subcommittee, I am Keith Rhodes, GAO's Chief Technologist and 
the Director of GAO's Center for Technology and Engineering.
    Although they are not with me at the table, I would like to 
acknowledge the study members, Jason Fong, Sushil Sharma, and 
James Tuitte.
    I am pleased to be here today to present our preliminary 
assessment of the plume modeling conducted by the Defense 
Department, and the Central Intelligence Agency, to determine 
the number of U.S. troops that might have been exposed to the 
release of chemical warfare agents during the first Gulf war of 
1990.
    We will be reporting the final results of this study at a 
later date. As you know, many of the approximately 700,000 
veterans of the first Gulf war have undiagnosed illnesses since 
the war's end in 1991.
    Some fear they are suffering from chronic disabling 
conditions because of wartime exposure to vaccines, as well as 
chemical warfare agents, pesticides, and other hazardous 
substances with known or suspected adverse health effects.
    Available bomb damage assessments during the war showed 
that of the 21 sites bombed in Iraq characterized by 
intelligence agencies as nuclear, biological or chemical 
facilities, 16 had been destroyed by bombing. Some of these 
sites were near the areas where U.S. troops were located.
    When the issue of the possible exposure of troops to low 
levels of chemical warfare agents was first raised during the 
summer of 1993, the DOD and the CIA concluded that no U.S. 
troops were exposed, because, No. 1, there were no forward-
deployed chemical warfare agent munitions; and No. 2, the 
clouds of chemical warfare agents or plumes from the bombing 
that destroyed the chemical facilities could not have reached 
the troops.
    DOD and CIA maintained this position until 1996 when it 
became known that U.S. troops destroyed a stockpile of chemical 
munitions after the first Gulf war in 1991, at a forward 
deployed site, Khamisiyah in Iraq. This discovery prompted 
several modeling efforts from 1996 through 2000 by DOD and CIA, 
to estimate the number of troops that might have been 
potentially exposed to chemical warfare agents.
    This modeling included field testing and modeling of 
bombing sites, as well as the number of U.S. troops exposed to 
the plume. The Department of Energy's Lawrence Livermore 
National Laboratory was also asked to conduct modeling.
    DOD and CIA created a composite of their own individual 
models and conducted additional plume modeling of the bombing 
sites at Al Muthanna, Muhammadiyat, and Ukhaydir.
    In addition, DOD used these models as the basis for their 
epidemiological studies regarding incidents of Gulf War 
Syndrome among U.S. troops returning from the first Gulf war. 
The dispersion of chemical agents was used to define the groups 
of people to be studied, those in theatre who were possibly 
exposed to chemical warfare agents, and those in theatre who 
were not.
    We disagree with the DOD and CIA conclusions for the 
following reasons: All modeling is limited. Models are not 
reality. They are, at best, an approximation of what will 
happen, or what did happen during a specific event. The 
validity of the model is a function of the data that forms the 
basis for the model.
    Thus, if you put garbage into it, you get garbage out of 
it. Thus, weak data inputs yield weak models, and from them 
weak analysis and conclusions. The DOD-CIA modeling efforts 
were weak in many ways: No. 1, meteorological data was 
incomplete and limited. For example, both the temperature at 
varying altitudes and over time was not complete; No. 2, the 
source term data, the data that defines how things reacted 
during the event and their potency were unknown and not 
reconstructed properly during field testing. For example, the 
purity of the agent was based on an UNSCOM report and was not 
consistent for all of the sites in question.
    One site had an agent purity estimated as high as 50 
percent, while another had a purity of only 15 percent, even 
through both sites were estimated to have the same agent which 
was manufactured at the same time.
    This limitation can be seen in that even though the same 
inputs were used for several model runs, the outputs differed 
significantly.
    Plume height--No 3. Plume height was arbitrarily selected 
to 10 meters, whereas independent field testing demonstrated 
that a single 1,000 pound bomb would create plume height in 
excess of 400 meters above the ground.
    No. 4, post-war field testing done at Dugway Proving Ground 
did not realistically simulate the actual conditions of 
bombings at any of the sites. The composite model that DOD and 
CIA made based on their earlier analyses, produced one pattern 
which removed the differences from the varying models, thus 
giving a much smaller range of differences between the possible 
plumes.
    The modeling and analysis executed by Livermore was 
discounted since it differed from the DOD and CIA analysis. 
Livermore did not agree for one main reason; they recognized 
that an atmospheric disturbance called a diffluence existed at 
the time of the demolition. This diffluence showed that the 
plume could have moved either to the north or south, or both to 
the north and south.
    As you can see on the story board here, the green area is 
the DOD composite model. The yellow and light yellow areas are 
the Livermore model. As you can see, because of that diffluence 
that went directly through the center of Khamisiyah, the 
Livermore model shows a wider range of dispersion than what the 
DOD models show.
    The problem is, there is no way of knowing exactly which 
one of these plumes is correct, or that both of them are 
correct, that the intersection of both of these models is 
correct, and therefore a much larger area was covered.
    Given these uncertainties in the modeling data, we conclude 
that: DOD cannot know who was and who was not exposed to any 
level of useful accuracy, since the method, model and data of 
the analysis are flawed, which calls into question DOD's 
conclusions based upon subsequent epidemiological studies, that 
those who were exposed had no higher rates of illnesses than 
those who were not exposed.
    Also, given the weaknesses in the data available for any 
further analysis, any further modeling efforts on this issue 
would not be any more accurate or helpful. We, therefore, 
recommend that the Congress direct the Secretary of Veterans 
Affairs to alter the assumptions regarding the Gulf War 
Syndrome to presume exposure, since many more veterans could 
possibly have been exposed than first estimated.
    Mr. Chairman, this concludes my statement. I will be happy 
to answer any questions you or members of the subcommittee may 
have.
    Mr. Turner [presiding]. Thank you, Mr. Rhodes.
    [The prepared statement of Mr. Rhodes follows:]

    [GRAPHIC] [TIFF OMITTED] T1396.004
    
    [GRAPHIC] [TIFF OMITTED] T1396.005
    
    [GRAPHIC] [TIFF OMITTED] T1396.006
    
    [GRAPHIC] [TIFF OMITTED] T1396.007
    
    [GRAPHIC] [TIFF OMITTED] T1396.008
    
    [GRAPHIC] [TIFF OMITTED] T1396.009
    
    [GRAPHIC] [TIFF OMITTED] T1396.010
    
    [GRAPHIC] [TIFF OMITTED] T1396.011
    
    [GRAPHIC] [TIFF OMITTED] T1396.012
    
    [GRAPHIC] [TIFF OMITTED] T1396.013
    
    [GRAPHIC] [TIFF OMITTED] T1396.014
    
    [GRAPHIC] [TIFF OMITTED] T1396.015
    
    [GRAPHIC] [TIFF OMITTED] T1396.016
    
    [GRAPHIC] [TIFF OMITTED] T1396.017
    
    [GRAPHIC] [TIFF OMITTED] T1396.018
    
    [GRAPHIC] [TIFF OMITTED] T1396.019
    
    [GRAPHIC] [TIFF OMITTED] T1396.020
    
    [GRAPHIC] [TIFF OMITTED] T1396.021
    
    [GRAPHIC] [TIFF OMITTED] T1396.022
    
    [GRAPHIC] [TIFF OMITTED] T1396.023
    
    [GRAPHIC] [TIFF OMITTED] T1396.024
    
    [GRAPHIC] [TIFF OMITTED] T1396.025
    
    [GRAPHIC] [TIFF OMITTED] T1396.026
    
    [GRAPHIC] [TIFF OMITTED] T1396.027
    
    [GRAPHIC] [TIFF OMITTED] T1396.028
    
    [GRAPHIC] [TIFF OMITTED] T1396.029
    
    [GRAPHIC] [TIFF OMITTED] T1396.030
    
    [GRAPHIC] [TIFF OMITTED] T1396.031
    
    [GRAPHIC] [TIFF OMITTED] T1396.032
    
    [GRAPHIC] [TIFF OMITTED] T1396.033
    
    [GRAPHIC] [TIFF OMITTED] T1396.034
    
    [GRAPHIC] [TIFF OMITTED] T1396.035
    
    [GRAPHIC] [TIFF OMITTED] T1396.036
    
    [GRAPHIC] [TIFF OMITTED] T1396.037
    
    [GRAPHIC] [TIFF OMITTED] T1396.038
    
    [GRAPHIC] [TIFF OMITTED] T1396.039
    
    [GRAPHIC] [TIFF OMITTED] T1396.040
    
    [GRAPHIC] [TIFF OMITTED] T1396.041
    
    [GRAPHIC] [TIFF OMITTED] T1396.042
    
    [GRAPHIC] [TIFF OMITTED] T1396.043
    
    [GRAPHIC] [TIFF OMITTED] T1396.044
    
    [GRAPHIC] [TIFF OMITTED] T1396.045
    
    [GRAPHIC] [TIFF OMITTED] T1396.046
    
    [GRAPHIC] [TIFF OMITTED] T1396.047
    
    [GRAPHIC] [TIFF OMITTED] T1396.048
    
    [GRAPHIC] [TIFF OMITTED] T1396.049
    
    [GRAPHIC] [TIFF OMITTED] T1396.050
    
    [GRAPHIC] [TIFF OMITTED] T1396.051
    
    [GRAPHIC] [TIFF OMITTED] T1396.052
    
    [GRAPHIC] [TIFF OMITTED] T1396.053
    
    [GRAPHIC] [TIFF OMITTED] T1396.054
    
    [GRAPHIC] [TIFF OMITTED] T1396.055
    
    [GRAPHIC] [TIFF OMITTED] T1396.056
    
    Mr. Turner. Dr. Johnson-Winegar.
    Dr. Johnson-Winegar. Mr. Chairman and committee members, I 
am honored to appear today before your committee to address 
your questions regarding the Department of Defense efforts to 
model chemical, biological, radiological and nuclear weapons 
effects.
    I am Dr. Anna Johnson-Winegar, the deputy assistant to the 
Secretary of Defense for Chemical and Biological Defense. In 
this role, I am responsible for the oversight and coordination 
of the Department of Defense chemical and biological defense 
programs.
    In addition, I have served as the accreditation authority 
within the Department for all common use chemical and 
biological defense models. In my testimony today, I will 
provide an overview of modeling to address some of the 
uncertainties inherent in all models. I request that my full 
written statement be incorporated into the record as it answers 
the specific questions posed to us in advance of today's 
hearing.
    All models and simulations are designed for specific 
purposes. Models are used for hazard prediction, risk analysis, 
operational decision support, virtual prototyping, weather 
forecasting and other purposes. In addition, models may be 
simple and easy to use, or complex and require expert users or 
indeed lie somewhere in between.
    No one model is suitable for all purposes. Conversely, only 
select models are appropriate for supporting specific analyses.
    Models are but a part of any analytical and decisionmaking 
process. While the selection of a model must be made in the 
context of the decision process that it will support, the 
actual efficacy of any model must begin with data or source 
terms.
    For a model to represent an event accurately, detailed 
knowledge about the event is essential. For chemical, 
biological, radiological and nuclear defense analysis, key 
information needed includes: Weather conditions, geographic 
conditions, type of threat agent, concentration and purity, 
state of agent, type of delivery systems and type of event.
    For example, dispersal from bulk storage as a result of 
counter-force operations, unconventional sources, toxic 
material accidents, etc.
    Uncertainty in these areas directly affects the accuracy of 
the model outputs. Once source terms are defined, models may 
calculate submunition and debris dispersal and propagation, and 
vapor, liquid, solid or aerosol transport and diffusion.
    The transport and diffusion of particles is only part of 
the overall equation. Transport and diffusion incorporates 
interaction of the agent with the atmosphere and with the 
surfaces on which the agents are dispersed. Once the agents are 
dispersed, analyses are required to determine the interactions 
between the agents and the environment, and perhaps most 
critically, to determine the interaction between agent and 
humans.
    It is not sufficient to determine the quantity of agent to 
which an individual is exposed, the actual effects on humans 
must be calculated. Effects may range from no observable 
effects to lethal effects and everything in between.
    These effects may be acute or chronic, and the response 
times may be immediate or delayed. A critical factor leading to 
the uncertainty in models is indeed the limited dosage data on 
human exposure to chemical or biological warfare agents.
    Effects of human exposure are primarily extrapolated from 
animal tests, along with analysis of some limited accidental 
exposures. All of these factors result in some degree of 
uncertainty in the output from all models.
    The role of models is to provide tolls to the analyst who 
then uses the output from the models to support decisionmaking. 
The analyst will incorporate risk assessment, sensitivity 
analysis and tradeoff analysis to account for uncertainty and 
to provide the most reasonable response germane to the question 
posed by the decisionmaker.
    Even though many of the same models used to model the 
activities related to the 1991 Gulf war are in use today, these 
models, in many ways, are the same in name only. There have 
been numerous advances in the capabilities of the various 
models. These advances have been integrated into the models 
currently in use to support hazard prediction, operational 
analysis and other activities.
    Each of these advances enhances the realism of the model, 
and enables the model to be used as a tool to provide a 
definitive estimate of the ground truth regarding the actual 
release of chemical or biological agents.
    In my written statement, I have provided a number of 
examples of the many enhancements of the models over the last 
10 years. In addition to enhancements of the models, there is a 
significant amount of data that must be measured. Not all data 
are essential for effective plume modeling.
    There is always a constant tradeoff in providing the most 
comprehensive data versus timely information versus high 
resolution.
    Also much of the data may be absent or estimated, because 
of natural variability that can only be described in a 
qualitative sense. Thus, even with perfect data, there will be 
uncertainties in an effective model because meteorology is 
inherently uncertain.
    Plume modeling and troop location data are linked in order 
to estimate potential effects of exposures on personnel and on 
the mission. Yet, the ability to model plumes to determine 
hazardous areas is not Affected by the location of the units. 
However, the ability to analyze possible exposures to service 
members in those units to the hazardous content of the plume 
often requires plume modeling in the absence of onsite testing.
    The separate data for plumes and troop location are tied 
together through our joint warning and reporting network, 
JWARN. Personnel and mission effects are then evaluated based 
upon the time-dependent hazard environment and the troop 
location in that environment.
    Currently JWARN troop location and plume data are tied 
together in a semiautomated manner. Planned upgrades over the 
next few years to JWARN will automate this process. The 
chemical-biological defense program has significantly increased 
its investment in the area of modeling and simulation over the 
last few years.
    Please be assured that the Department takes this very 
seriously. We understand our responsibility to provide the most 
accurate information possible related to transport and 
diffusion of these types of agents. We are indeed working very 
closely with other programs in the Department of Defense as 
well as with the other Federal agencies.
    Thank you for the opportunity to address these questions, 
and I remain available to try to answer any further questions 
or concerns that the committee may have.
    Mr. Turner. Thank you, Dr. Johnson-Winegar.
    [The prepared statement of Dr. Johnson-Winegar follows:]

    [GRAPHIC] [TIFF OMITTED] T1396.057
    
    [GRAPHIC] [TIFF OMITTED] T1396.058
    
    [GRAPHIC] [TIFF OMITTED] T1396.059
    
    [GRAPHIC] [TIFF OMITTED] T1396.060
    
    [GRAPHIC] [TIFF OMITTED] T1396.061
    
    [GRAPHIC] [TIFF OMITTED] T1396.062
    
    [GRAPHIC] [TIFF OMITTED] T1396.063
    
    [GRAPHIC] [TIFF OMITTED] T1396.064
    
    [GRAPHIC] [TIFF OMITTED] T1396.065
    
    [GRAPHIC] [TIFF OMITTED] T1396.066
    
    [GRAPHIC] [TIFF OMITTED] T1396.067
    
    [GRAPHIC] [TIFF OMITTED] T1396.068
    
    [GRAPHIC] [TIFF OMITTED] T1396.069
    
    [GRAPHIC] [TIFF OMITTED] T1396.070
    
    Mr. Turner. Dr. Ermak.
    Dr. Ermak. Mr. Chairman, and members of the subcommittee, 
thank you for the opportunity to appear before you today. My 
name is Don Ermak, and I lead the National Atmospheric Release 
Advisory Center [NARAC], at Lawrence Livermore National 
Laboratory.
    The opinions that I present today represent my views, and I 
would like to focus on plume prediction and the development 
that is needed to address current threats to national security.
    NARAC calculations for the Khamisiyah incident. NARAC is a 
Department of Energy and Department of Homeland Security 
operational support and resource center for plume modeling. Its 
mission is to provide timely and credible advisories to 
emergency managers for hazardous releases to the atmosphere.
    In October 1996, the CIA asked NARAC to calculate the 
atmospheric dispersion of Sarin resulting from U.S. demolition 
activities in March 1991 at the Khamisiyah munitions storage 
facility.
    We conducted three hypothetical release scenarios as 
specified by the CIA. In November, at the request of the DOD, 
Dr. Michael Bradley presented the NARAC results to an IDA panel 
on low-level exposure to chemical agents. At that meeting, we 
were asked to do additional simulations, and to present the 
results in February, which we did.
    NARAC was not asked to participate in further studies. At 
that time, we were not convinced that all paths to understand 
the event had been exhausted. However, since then, several 
other attempts have been made.
    Unfortunately, both the weather observations and the source 
of data appear to be inadequate for any model to provide a 
single definitive simulation. It is not clear to us that 
further analysis are warranted.
    Current challenges. Recent terrorist events have heightened 
national concern over urban terrorism and the release of 
airborne, nuclear, biological and chemical agents. In response 
to these and other concerns, we have expanded our sources of 
real-time and forecast weather data, enhanced our modeling 
capabilities to treat biological and chemical agent releases, 
and the bulk effects of urban areas, and have developed 
Internet and Web-based communications for easy user access to 
NARAC.
    We have also developed a state-of-the-science building 
scale model that simulates flow and dispersion around buildings 
for planning and special events. More work is needed. Both new 
capabilities and the expanded application of existing 
capabilities are needed to address this critical national 
security concern.
    First, enhanced meteorological data networks. Atmospheric 
dispersion models are powerful tools. However, all dispersion 
models require high quality weather observations. More weather 
observation locations are needed for models to accurately 
predict plumes in urban areas.
    Of particular note is the need for upper level air 
observations. Second, urban dispersion modeling. High fidelity, 
building to urban to regional scale dispersion simulations are 
essential for vulnerability studies, risk assessments, 
attribution and intelligence applications.
    In addition, these models can answer important questions 
concerning building infiltration, command post siting, and 
evacuation routes for emergency response.
    Third. Studies of atmospheric transitions. Many 
metropolitan areas are within 20 miles of an ocean or large 
body of water. Land-sea breezes change the direction and speed 
of the winds throughout the course of a day. Additional 
meteorological observations and improved fine scale weather 
prediction models are needed to provide accurate and reliable 
predictions in the coastal environment.
    Fourth. Model evaluation. We see several key elements in 
model evaluation. Analytic comparisons, comparisons with field 
experiments, operational testing to evaluate robustness, and 
open literature publication and public availability to allow 
for scrutiny by the scientific and user communities.
    While it is not practical to verify the models under all 
conditions, we strongly support continued field programs 
focused on the issues discussed above.
    Fifth. A systems approach. In addition to data 
assimilation, weather prediction and plume dispersion models, 
an effective response capability needs to include dependable 
voice and data communications, rapid high-volume atmospheric 
data collection and extensive data bases of terrain, maps, 
population and health effects.
    Of critical importance are situation awareness tools that 
provide emergency managers with a clear picture of the hazard. 
Event reconstruction capabilities that integrate observational 
data with prediction models are needed to estimate poorly known 
sources. And, finally, a highly trained multi-disciplinary 
staff is needed for reach-back during events.
    The development of such a capability is being explored by 
the DOE and DHS Link program. The objective is to demonstrate 
the capability for providing local government agencies with 
NARAC capabilities in a manner that can be seamlessly 
integrated with appropriate State and Federal agency support. 
We are currently working with the cities of Seattle and New 
York.
    In closing, let me assure you that we at NARAC are 
dedicated to the state-of-the-science plume prediction and 
emergency response support to meet the Nation's security needs. 
Thank you.
    Mr. Turner. Thank you.
    [The prepared statement of Dr. Ermak follows:]

    [GRAPHIC] [TIFF OMITTED] T1396.071
    
    [GRAPHIC] [TIFF OMITTED] T1396.072
    
    [GRAPHIC] [TIFF OMITTED] T1396.073
    
    [GRAPHIC] [TIFF OMITTED] T1396.074
    
    [GRAPHIC] [TIFF OMITTED] T1396.075
    
    [GRAPHIC] [TIFF OMITTED] T1396.076
    
    [GRAPHIC] [TIFF OMITTED] T1396.077
    
    [GRAPHIC] [TIFF OMITTED] T1396.078
    
    [GRAPHIC] [TIFF OMITTED] T1396.079
    
    Mr. Turner. Mr. Hicks.
    Mr. Hicks. Good afternoon, Mr. Chairman, and members of the 
subcommittee. My name is Bruce Hicks, and I am Director of the 
Air Resources Laboratory of the National Oceanic and 
Atmospheric Administration [NOAA]. I have been actively 
involved in studies of the transport and diffusion of 
pollutants in the atmosphere for more than 40 years, with 
research experience in Australia and at several U.S. 
laboratories. I have been with NOAA since 1980.
    I recently served as the cochairman of the joint action 
group for the selection and evaluation of atmospheric transport 
and diffusion models set up by the office of the Federal 
coordinator for meteorology.
    I have been asked to present some views regarding the 
current state of the science in the modeling of atmospheric 
dispersion. It is my pleasure to do so. I would like to show 
three diagrams later as I speak.
    It is a major part of the mission statement of NOAA to 
provided forecasts to protect the public. Forecasts of 
atmospheric dispersion are among the capabilities that we 
provide.
    The Chernobyl nuclear accident is an example where 
dispersion models were used real-time for an unfolding 
emergency situation. The results showed that many dispersion 
forecasts were quite deficient. The World Meteorological 
Organization concluded that there was need for a more organized 
provision of dispersion forecasts in the future, and hence set 
up a small network of international recognized dispersion 
forecast providers.
    There are now seven of these in France, England, Canada, 
Russia, China, Australia, and in the USA.
    In practice, Montreal serves as a back-up to the U.S. 
capability in NOAA and vice versa. There are 122 weather 
forecast officers of the National Weather Service nationwide. 
In the event of an incident requiring the forecast of 
dispersion, each of those centers is prepared to provide 
dispersion predictions out to at least 2 days.
    The accuracy of the dispersion forecasts depends on the 
accuracy with which the meteorological wind fields are known. 
Operational weather forecasts guidance is available at 12 
kilometer resolution. And the weather service forecasters are 
now beginning to provide rude forecast windshields at even 
higher resolution.
    The model we use, HYSPLIT, is operationally integrated with 
the Weather Service's highest resolution weather prediction 
models, and takes advantage of greater resolution, both spatial 
and temporal, within the model stream at a data density higher 
than is generally practical for rapid external distribution.
    Dispersion predictions for selected locations across the 
Nation are made with updated weather forecast data four times 
each day. For emergency events or preparations, dispersion 
predictions are run, on request, at 12-kilometer resolution and 
results are generally available within 15 minutes.
    4-kilometer resolution predictions can also be run on 
demand. The model, HYSPLIT, is also run on remote computer 
systems. But, these remotely run applications rely on reduced 
resolution weather data to drive the dispersion calculation. 
All of the weather forecast officers have access to both kinds 
of product.
    The course of product is also available via the Internet to 
users to who are registered by our laboratory. This is known as 
the real-time environmental applications and display system, 
READY, which is used routinely by over 1,500 registered users.
    The READY system brings together dispersion models, display 
programs, and forecast programs generated over many years in a 
form that can be used by anyone. The products are used, for 
example, to guide response activities following industrial 
accidents and forest fires.
    The data have been used by every long distance manned 
balloon venture so far. The READY System is widely known and 
routinely employed. The models that now make be READY were 
central in the activities addressing the Kuwait oil fires back 
in 1990 and 1991.
    To us, the Khamisiyah experience was quite revealing and is 
worthy of some direct attention. In their scrutiny of the 
subject, the Office of the Special Assistant for Gulf War 
Illness, elected to use a small number of dispersion models, 
mainly from within the DOD system. There were indeed very few 
meteorological observations available, and hence, the 
dispersion codes were driven by exceedingly sparse and 
sometimes questionable information.
    To us it is not surprising that the dispersion systems 
yielded different answers. Each one of these answers 
represented a good approach to the problem. There was no way to 
weigh or order these alternative depictions of the plume from 
Khamisiyah.
    The community has now adopted the concept of ensemble 
modeling, in which many models are used to address the 
situation, and the answers are derived from analysis of all of 
their products. This was much like what was done for 
Khamisiyah, but on a larger scale.
    In North America, we are not short on data, although we 
still have need to learn how to use the available information 
optimally. The shortcomings that caused the dismay about 
Khamisiyah should not be seen as a basis for concern 
necessarily in North American situations. I should point out 
that among other products, the READY system maintains a 
continuously updated plume forecast for every nuclear power 
plant installation in North America.
    In the event of a release of radioactivity from any nuclear 
power plant, there is no need to start a dispersion forecast, 
it is always immediately available. All that is needed is a 
password and access to the relevant READY product.
    So far, I have emphasized the long-range aspect of the 
problem. Much of the focus of present concern is on urban 
cities, and urban areas and cities. NOAA, in partnership with 
EPA, provides a local dispersion capability with the CAMEO/
ALOHA system. CAMEO is the Computer Aided Management of 
Emergency Operations System. The neofield atmospheric 
dispersion model, provided in conjunction with CAMEO is ALOHA, 
the air relocations with hazardous atmospheres model.
    First responders and emergency planners use CAMEO to plan 
for and to respond to chemical emergencies. More than 30,000 
copies of this model system have been distributed to users 
across the country in the last year.
    Over 1,000 local responders will receive training during 
the next year. It is for cities and urban areas that the 
greatest challenge exists. The monitoring stations used by the 
Weather Service at this time are typically located at airports, 
but the area of main concern is usually quite distant from the 
airports.
    May I have the first visual, please? Is it possible? Here 
you will see an example of a dispersion product, showing spread 
of materials from Ohio across the United States. This product 
is based upon weather--is based upon the results of actual 
release of material. This is not a forecast. This is what the 
plume actually looks like. That is what we are trying to 
forecast.
    In practice, though, the information that we use when we 
try to forecast that comes from the weather forecast officers 
and from largely the airports in areas.
    It is the wind fields that determine the released material 
and where it will drift to. And it is the atmospheric 
turbulence that controls the rate at which dilution occurs. 
Both are strongly affected by the presence of buildings or 
other structures.
    This is large scale. Now I want to talk about the smaller 
scale. The Nation has many atmospheric dispersion model that 
purport to predict the dispersion of hazardous materials 
released into the urban atmosphere. The capabilities are 
widespread across the Federal agencies.
    Every one of these systems has some special quality that 
makes it unique. The trick now facing the atmospheric 
dispersion community is to determine which subset of the many 
dispersion systems is best suited to the latest challenges.
    In a recent report, the Office of the Federal Coordinator 
concludes that there are 29 modeling systems running 24 by 7 
within the Federal system. Of these, seven systems are used 
nationwide, including HYSPLIT NOAA.
    Recent field studies in Salt Lake City, for example, have 
yielded a lot of new information. However, we do not yet know 
how to apply the results so that they may be applicable for 
some specific urban area to another, with confidence.
    Consequently, there is a strong need to obtain relevant 
data. This is the basis for the design of what we refer to as 
DCNet, a program to provide Washington with the best possible 
basis for dispersion computation as is needed for both planning 
and possible response.
    The problem we face is complex. The windows within a city 
sometimes bear little resemblance to those in the surrounding 
countryside, as I have already said. For small street level 
releases of pollutants, these local scale conditions are 
dominant, especially within the first minutes to hours, until 
entrainment above the buildings is dominant.
    The presence of buildings and the street canyons separating 
them often causes winds that are almost random, exceedingly 
difficult to predict or even describe.
    The flow above the urban canopy is far more describable in 
terms of larger scale meteorology. It is convenient to think in 
terms of two regimes, the street canyon flows beneath the urban 
canopy and the skimming flow above it.
    Washington presents an excellent test bed for studies, 
because the urban canopy is well defined by the height 
constraints of the buildings. New York, for example, presents 
an opposite extreme.
    In New York, many buildings are not only very high, but 
their height is quite variable. Thus, there are two reasons for 
focusing on the Washington metropolitan area.
    Mr. Turner. Mr. Hicks, are you near conclusion?
    Mr. Hicks. Thank you. I would like to show you the next 
slide, which shows the array of sites presently deployed in the 
Washington area. There are approximately 13 locations where 
special instrumentation is being deployed.
    And the last slide shows you the main point that I would 
like to reach. It shows you the window roses. It is a depiction 
of the wind directions from different locations across the 
Washington area, which has been shown by the DCNet operation. 
They are quite different.
    Mr. Chairman, I would like to close with that. That 
concludes my testimony. Thank you for the opportunity. I would 
be happy to respond to any questions that the subcommittee 
might have.
    Mr. Turner. Thank you, Mr. Hicks.
    [The prepared statement of Mr. Hicks follows:]

    [GRAPHIC] [TIFF OMITTED] T1396.080
    
    [GRAPHIC] [TIFF OMITTED] T1396.081
    
    [GRAPHIC] [TIFF OMITTED] T1396.082
    
    [GRAPHIC] [TIFF OMITTED] T1396.083
    
    [GRAPHIC] [TIFF OMITTED] T1396.084
    
    [GRAPHIC] [TIFF OMITTED] T1396.085
    
    [GRAPHIC] [TIFF OMITTED] T1396.086
    
    [GRAPHIC] [TIFF OMITTED] T1396.087
    
    Mr. Turner. Dr. Barron.
    Dr. Barron. Good afternoon. My name is Eric Barron. As 
Chair of the National Research Council's Board on Atmospheric 
Sciences and Climate, I am here today to discuss the Board's 
new report, entitled ``Tracking and Predicting the Atmospheric 
Dispersion of Hazardous Material Releases: Implications for 
Homeland Security.''
    There are three phases to addressing deliberate release of 
hazardous materials, such as chemical, biological or nuclear 
agents. Preparedness, response, and recovery and analysis.
    The atmospheric sciences contributes to all three. In the 
preparedness phase, we can enable risk assessment, improve 
training exercises and aid in evaluating outcomes associated 
with potential sites of hazardous release. The preparation for 
the Salt Like City Olympics is a good example of this mode of 
operation.
    In recovery and analysis, atmospheric models and 
observations can be used to examine exposure levels. Such 
assessments were utilized extensively following both Chernobyl 
and September 11th in both real-time and in terms of recovery.
    Response is a much greater challenge, because time is of 
the essence and vulnerable regions such as major cities present 
special challenges. In every one of those phases, improvements 
in capability are warranted, but it is particularly in the area 
of response that the needs of first responders and emergency 
managers do not seem to be well satisfied by existing 
capabilities.
    Our capacity to meet these challenges rests on three 
interconnected elements: Atmospheric dispersion models that 
predict the path and spread of hazardous agents, observations 
of the plume and local meteorological conditions, and effective 
coordination among the relevant atmospheric science and 
emergency response communities.
    The committee recommends that we establish a nationally 
coordinated effort for the support and evaluation of existing 
models and development of new modeling approaches. The Office 
of the Federal Coordination for Meteorology has taken some 
important initial steps in this regard.
    As a part of this effort, the report concludes that we must 
focus on operational and specifically urban use of these 
models, develop model solutions that specifically quantify 
confidence levels and the nature of variability of the 
predictions, enhance our ability to assimilate meteorological, 
primarily wind, temperature and moisture, and CBM sensor data 
into models, conduct urban field programs and wind tunnel 
simulations to better evaluate and to better develop models, 
and to focus on rigorous and independent model intercomparisons 
and evaluations.
    In terms of observations, the committee recommends that we 
conduct comprehensive surveys of existing observational 
networks and work to improve those networks, especially around 
key vulnerable areas. Here the most important points are to 
improve our ability to identify the source and the plume, 
characterize low-level winds, characterize the depth and 
intensity of atmospheric turbulence, and identify areas of 
potential degradation and dry or wet deposition of the harmful 
agents, to explore supplementing existing radar network with 
short-wave length radars that enable better meteorological 
observations and better identification of the plumes.
    To continue to develop airborne and surface mobile 
observation platforms with a focus on rapid deployment and 
accessibility, and to conduct field programs with the objective 
of using observations to test and modify dispersion and missile 
scale transport models.
    The committee is also concerned that emergency managers 
need a more realistic understanding of the uncertainties 
associated with dispersion prediction, and the atmospheric 
sciences community should have a better understanding of the 
needs of responders. In particular, the committee recommends 
Table Top event simulation exercises convened on a regular 
basis to bring together response teams and members of the 
atmospheric sciences community to help establish and exercise a 
common set of data interface and decision support protocols.
    And, two, a more carefully crafted management strategy with 
a strong center of coordination and clear lines of 
responsibility. We suggest a single point of contact to connect 
emergency responders to appropriate modeling centers for 
immediate assistance.
    In at least one urban area, a fully operational dispersion 
tracking and forecasting system should be established. This 
should be a comprehensive system designed as a test bed for 
understanding and improving our capabilities, and providing the 
basis for a much broader national implementation.
    As a final point, it should be emphasized that robust 
atmospheric observing systems and high resolution atmospheric 
modeling systems will be used for many other important 
purposes, to support severe weather warnings, for air quality 
forecasting, and of course for tracking the accidental release 
of some hazardous material.
    Such multiple uses will help justify costs and ensure that 
the systems are regularly maintained and evaluated. I would 
like to thank the subcommittee for this invitation to testify, 
and I would be happy to answer any questions.
    Mr. Turner. Thank you, Dr. Barron.
    [The prepared statement of Dr. Barron follows:]

    [GRAPHIC] [TIFF OMITTED] T1396.088
    
    [GRAPHIC] [TIFF OMITTED] T1396.089
    
    [GRAPHIC] [TIFF OMITTED] T1396.090
    
    [GRAPHIC] [TIFF OMITTED] T1396.091
    
    Mr. Turner. Dr. Hanna.
    Dr. Hanna. I would like to thank the subcommittee for 
asking me to testify. My name is Steven Hanna, and I am with 
the Harvard School of Public Health. And I represent a person 
who has done research on turbulence and dispersion modeling for 
many years.
    I am representing myself, so my opinions are my own, based 
on the science. I am probably the only person sitting here that 
doesn't have a staff backing me up ready to provide things.
    I have looked at a lot of government dispersion models over 
the years. That is probably the reason I have been asked to 
testify, including EPA, Department of Defense, NOAA, and other 
types of modeling systems.
    I was the chair of a three-member peer review committee of 
the Khamisiyah modeling exercise for several years. And I must 
say, our conclusions about how good the exercise was are 
somewhere in between the two speakers on the other end of the 
table.
    I would like to first review some fundamental facts about 
transport and dispersion models. One has been mentioned before, 
is that much of the history of this field comes from chemical, 
biological agents from needs in World War 1.
    One interesting aspect of them, is you run them in a 
forward or backward mode. If you know what the source is, you 
can calculate what is going to happen to people.
    On the other hand, if you don't know where the source was, 
you can use observations of concentrations in order to try to 
triangulate back to where the source might have been.
    Another fundamental fact is also substances move in a 
similar manner, the chemical agents, biological agents, other 
types of tracers are dispersed through the atmosphere 
similarly, and you can use the same types of models.
    The difference between emergency response and other types 
of more routine models is that the emergency response has to 
run fast, and needs to have capability of bringing data into 
the system.
    Another difference between chemical agents and biological 
agents in the way they can be run and interpreted, because 
there is an immediate effect of a chemical agent, so you can do 
emergency response modeling, but with biological agents, it 
requires 2 weeks later before people start showing up at 
emergency rooms. But, you can still do planning studies.
    The uncertainties have been addressed by the others, and 
you can think of it in terms of weather forecasts. We all know 
how certain weather forecasts are, and the same thing applies 
to transfer and dispersion models, because there the material 
is moving in the atmosphere.
    I would like to point out that over the past 10 years there 
have been great improvements to DOD, DOE and other dispersion 
modeling systems so that many of them are now capable of 
modeling things with state-of-the-art science.
    A couple of major issues. I feel that the government 
assessments have been ignoring the many valuable models 
available from the industries. The chemical processing plants 
and oil refinery industries have developed many very good 
models that I don't see being used or considered.
    Another issue is I see a--it is quite unclear on who runs 
which model when we have several agencies who are running 
models for emergency response. And I have seen those written 
down. But I have a hard time myself deciding this, and I think 
there needs to be more definition.
    Another issue is, I believe with some of the models that 
are out there for use by the general public or by the military, 
that we need more consistency in user guidance. I see 100 
different users getting different answers when they run the 
same model against the same scenario.
    We need better field tests. Most of our field tests so far 
are what you would call fair weather. When we do an experiment, 
if it starts raining, the experimentalists pack up and go back 
to their hotel rooms. And real releases are just as likely to 
be during rain or when a front is going through. So we need 
more comprehensive studies.
    As for urban city areas, there is much discussion about the 
variability in the city. But, on the positive side, because of 
all of the buildings, there is a lot of mixing, and we find 
that in some aspects, especially at moderate distances, you can 
do quite well with modeling in cities.
    However, you do need the local observations because you 
obviously need to know which way the wind is blowing as the 
primary determinant.
    And my final comment is on the Gulf war. I believe that it 
was a reasonable program, the results were reasonable. However, 
seemed in many cases to be a compromise, and instead of being a 
long-term basic research effort, it seemed to be carried out in 
short bursts of 2-week subtasks rather than over a longer-term 
period.
    Thank you and I would be willing to answer further 
questions.
    Mr. Turner. Thank you, Dr. Hanna.
    [The prepared statement of Dr. Hanna follows:]

    [GRAPHIC] [TIFF OMITTED] T1396.092
    
    [GRAPHIC] [TIFF OMITTED] T1396.093
    
    [GRAPHIC] [TIFF OMITTED] T1396.094
    
    [GRAPHIC] [TIFF OMITTED] T1396.095
    
    [GRAPHIC] [TIFF OMITTED] T1396.096
    
    [GRAPHIC] [TIFF OMITTED] T1396.097
    
    [GRAPHIC] [TIFF OMITTED] T1396.098
    
    [GRAPHIC] [TIFF OMITTED] T1396.099
    
    Mr. Turner [presiding]. We will now turn to a series of 
questions--before we turn to questions, we actually want to ask 
unanimous consent that all members of the subcommittee be 
permitted to place any opening statement in the record, and 
that the record remain open for 3 days for that purpose.
    Without objection, it is so ordered. Also, I ask for 
further unanimous consent that all witnesses be permitted to 
include their written statements in the record. Without 
objection, it is so ordered.
    We will go first to questions from our chairman, Chairman 
Shays.
    Mr. Shays. Thank you. Also I have a unanimous consent, Mr. 
Chairman. The GAO request for this work was submitted jointly 
by this committee and Senator Robert Byrd of West Virginia, and 
I ask unanimous consent that a statement by Senator Byrd be 
included in the record.
    Mr. Turner. Without objection, so ordered.
    [The prepared statement of Hon. Robert C. Byrd follows:]

    [GRAPHIC] [TIFF OMITTED] T1396.100
    
    [GRAPHIC] [TIFF OMITTED] T1396.101
    
    [GRAPHIC] [TIFF OMITTED] T1396.102
    
    Mr. Shays. I thank all of our six witnesses for their 
attendance. In preparation for this hearing, I really wrestled 
with whether this is an art or a science. So I would like, if 
an art is 1 and a science is 10, where on the scale are we? I 
would like each of you to tell me. Is the projection of a plume 
an art or a science? Is it a 1 or a 10 or somewhere in between?
    Mr. Rhodes.
    Mr. Rhodes. Thank you, Mr. Chairman.
    I guess the way I would characterize it is that it is a 
genuine science. The interpretation of what you do with what 
comes out of the model can be more art even though the 
underlying science is a 10. It is real science. It is 
mathematics.
    Mr. Shays. But when you are done, what do I have? Do I have 
more science or more art?
    Mr. Rhodes. Depends on what you were able to put into the 
model. If you are able----
    Mr. Shays. Under the best conditions, what do I have?
    Mr. Rhodes. You still don't have reality. I mean, you still 
do not have reality today. So you still have an estimate, so it 
is not going to be----
    Mr. Shays. It is not going to be a mathematical certainty.
    Mr. Rhodes. It is not going to be a mathematical certainty. 
It will be mathematical, but it will not be a certainty.
    Mr. Shays. Maybe I need to ask the question differently. So 
you have succeeded in not giving me a number. You have 
basically told me it is a science, but there is no certainty, 
and the outcome is like an art, and it is only as good as what 
you put in. I understand that part. I am going to come back.
    Dr. Winegar, what do you think?
    Dr. Johnson-Winegar. I certainly agree in a qualitative 
sense that it is more of a science than an art. I would 
certainly want to characterize it as a science that is rapidly 
improving, and that any answer given today would certainly only 
be a snapshot in time as to how much of a science it is in 
comparison, for example, of what was done 10 years ago or, even 
more to the point, what we will be able to do 10 years from 
today.
    So I clearly put modeling and simulation in the area of 
science.
    Mr. Shays. Dr. Ermak.
    Dr. Ermak. I would agree with the past statements that the 
study of the atmosphere is a science. Like all sciences, there 
are things that are unknown. So I think that where the--perhaps 
the shift between science and art comes in is in the 
application of that, and that probably varies, depending on the 
application. In emergency response, where you cannot bring in--
where time limits our response, and all the resources of the 
research, and you do not have all the time to make this type of 
a prediction, then I think there is more art.
    Mr. Turner. Mr. Hicks.
    Mr. Hicks. In my judgment, it is a 7, and the science is 
trying to turn it into an 8.
    Mr. Shays. The last part I missed.
    Mr. Hicks. The research is trying to turn it into an 8.
    Mr. Shays. OK. Thank you.
    Dr. Barron. It is science on its way to becoming even 
better science. If you took the best case of extensive 
observations, a great deal of time to begin to do multiple 
simulations, create ensembles, understand something about 
uncertainties, I would say it was in the 7 to 8 range, in 
talking about confidence, not whether it is a science or not. 
In the worst case of very poor observations and the need for 
immediate response, then the confidence goes way down.
    Mr. Shays. OK. Dr. Hanna.
    Dr. Hanna. I would give it an 8 if you have a lot of data. 
If you are in a place like Khamisiyah, I would bring it down to 
a 6 or so. But in all cases, there is a lot of uncertainty.
    Mr. Shays. Thank you very much for your response. 
Remembering that, I am looking at three shades of color, green, 
and an olive green and a more yellow. They are all plumes, 
projections of plumes, correct?
    Mr. Rhodes. Yes, they are.
    Mr. Shays. They are the same incident, correct?
    Mr. Rhodes. Yes, they are.
    Mr. Shays. Defense is green?
    Mr. Rhodes. Deep green, yes.
    Mr. Shays. Deep green.
    Livermore is the more yellow.
    Mr. Rhodes. Yes, and the olive green around. The Defense 
composite is the one that drives down into Saudi Arabia, and 
Livermore is the one that moves up to Iran.
    Mr. Shays. Actually the yellow is blown out by the dark 
green, because the Khamisiyah, as I look at it, is at the very 
top.
    Mr. Rhodes. Yes. If you look at the highest point of the 
green composite, that is the side of Khamisiyah.
    Mr. Shays. The only certainty, at least with these two, is 
that the plume went south rather than north?
    Mr. Rhodes. Yes, initially. But as you can see from the 
Livermore model, it does turn and then start to move north.
    Mr. Shays. Correct. Originally started down.
    Mr. Rhodes. Yes.
    Mr. Shays. This may seem less of a focus for our other 
witnesses, but we have had 10 years of hearings on the whole 
issue of Gulf war illnesses in this committee, and we didn't 
know about Khamisiyah until we had a witness who actually had a 
video of blowing it up, and the Defense Department heard of our 
hearing that we were going to have the next week on Tuesday, 
notified the press at 12 noon on Friday that there would be a 4 
o'clock press hearing in which they said our troops were 
exposed to defensive chemical weapons, because they had denied 
that our troops had ever been exposed, and then we were getting 
in the word game of offense/defense.
    What is important to me here in this issue is that I 
believe that Defense basically looked at the soldiers who were 
under the dark green; is that correct?
    Mr. Rhodes. That is true.
    Mr. Shays. We made some presumption that anyone who could 
confirm that they were in the green, dark green, area had some 
exposure to chemical defensive exposure to chemicals; is that 
correct?
    Mr. Rhodes. That is true.
    Mr. Shays. If, in fact, we use the Livermore model, then 
all of the assumptions about who was exposed and who wasn't 
exposed become very different, correct?
    Mr. Rhodes. Yes, sir.
    Mr. Shays. So it is your recommendation, and I don't want 
it to get lost, but it is your recommendation that what 
happened? I want you to repeat it. It is on page 5. You say, 
``We, therefore, recommend''----
    Mr. Rhodes. We, therefore, recommend that the Congress 
direct the Secretary of Veterans Affairs to presume exposure, 
that those in theater are presumed exposed, because outside of 
that green area, those people in the area that the Livermore 
model shows should be exposed, and therefore, we are 
presuming--we are recommending that you direct the Secretary of 
Veterans Affairs to presume exposure of all veterans in that 
area, the total area, not just the composite area.
    Mr. Shays. Now, for the purpose of this hearing, we have 
two interests, I have at least two interests, but one is we 
still have to care about our veterans who were in the first 
Gulf war, because I felt shortly after the war DOD and the 
Department of Veterans Affairs didn't care enough about them. I 
take it they care now about them. So that is one issue.
    The other issue is just understanding this whole art to 
science, which I understand is more of a science, depending on 
the data, and understanding its impact in any future war and 
battle, and also to understand it domestically, because there 
is an absolute certainly that someday, somewhere in the United 
States, American civilians and all who are here in the United 
States in that particular area will be exposed to some 
chemical, biological, radioactive, nuclear, whatever. So it is 
very important. In other words, the work you do is very 
important, and some of you spend your mornings, noons, and 
nights thinking about this one issue. Thank you for doing that.
    But if you could just deal with this issue here right now, 
I would like you, Dr. Johnson-Winegar, to tell me how you react 
to what Mr. Rhodes has said and Dr. Ermak, as well as, Mr. 
Hicks and Dr. Hanna, if you would react to that.
    Let me just say also, Dr. Johnson-Winegar, one, I 
appreciate you participating. I have said in the past, but it 
is to your credit that you are so into participating in a 
larger panel because it makes us have better dialog, and we do 
thank you for that. Also I want to say that your statement 
clearly was comprehensive. It would have probably taken you 20 
minutes to go through, so I want to thank you that you did not 
do that, but I am also thankful that you took each question we 
asked and responded to it in a very thorough way, and we 
appreciate that.
    Having said that, could you react to what Mr. Rhodes said?
    Dr. Johnson-Winegar. Certainly. Let me just make two 
comments. First of all, the area that is shown in green, which 
is referred to as the DOD estimate, is, in fact, really a 
composite of information that was generated from using a number 
of different model systems, and the DOD did call upon IDA, the 
Institute for Defense Analysis, an independent organization, to 
review that data. The data was then subsequently reviewed yet 
again by scientists who are eminent in the field, some of whom 
are here today. So we have peer review accreditation of the 
work that was done.
    With regard to the apparent discrepancy between the DOD 
prediction and that run by the Lawrence Livermore model, I 
believe that the real answer is in what is being done with that 
information, and it is certainly my understanding that veterans 
are being treated based on symptomatology and not based solely 
on where they were geographically located; in other words, 
whether they were ``in the plume'' or ``not in the plume.'' So 
I think that the bottom line is we certainly appreciate your 
concerns, and I want to reinforce the concerns from the 
Department of Defense for all those veterans. Clearly, we agree 
with the premise that they should be treated based on potential 
exposure.
    Mr. Shays. I will come back, Dr. Johnson-Winegar. Thank 
you.
    Dr. Ermak.
    Dr. Ermak. Yes. When I look at this chart, I see this as an 
example of the uncertainty that often results from dispersion 
models and, in particular, the large uncertainty that can 
result when there is inadequate initial data on which to make 
the dispersion calculation. I will stop there.
    Mr. Hicks. Yes. My comments on this are colored by the fact 
that I was a member of the team that reviewed the Department of 
Defense work to start off with.
    Mr. Shays. I consider that helpful. I mean, thank you for 
saying that, but now react.
    Mr. Hicks. We delved very deeply into the assumptions that 
were made in that analysis, and I have not had the opportunity 
to do the same thing to the Lawrence Livermore analysis. I 
can't imagine what I would find if I were to do so, but at the 
moment I would say that I agree with Dr. Ermak that these are 
examples of how the plume forecasts are at the mercy of the 
assumptions that you make.
    Mr. Shays. OK. Thank you.
    Dr. Barron. This wasn't a specific part of the National 
Research Council's investigation, so the report doesn't--
whatever I say is outside the nature of that report. But I 
think that is the perfect answer. When you have inadequate 
observations, especially to initiate models, you can expect 
widely different simulations from the plume models.
    Dr. Hanna. This is exactly what we see in any sort of 
modeling exercise like this, and it makes me wonder why stop at 
five models? If we put 70 models up there, we would probably 
cover the entire 360 degrees.
    Mr. Shays. But having said that, then for me as a 
policymaker who has to be concerned about veterans, I sent to--
along with others in the first Gulf war, I look at that and I 
say that we can't be any more certain that the DOD model, based 
on a number of models put together, or the Livermore, is more 
accurate, and, therefore, it would strike me that we would have 
to give the presumption to the veteran that they were, in fact, 
exposed.
    Dr. Hanna. Well, I would interpret in a probability sense 
that there is a higher probability of people being affected in 
the middle of that group of plumes and lower probability at the 
outside of it.
    Mr. Shays. Right. The problem is--and I will get to 
questions about plumes of chemical, biological or radiological, 
and I will have some questions there--but the problem that 
differs for you and then for us is that men and women risk 
their lives in battle, and we don't really know where the plume 
was. That is really what--and yet we are trying to say we do, 
and we give a presumption if you are under the green, but if 
you are not under the green, you don't get the presumption. So 
that is a huge, a huge issue, at least for the committee.
    Thank you, Mr. Chairman. I will have some more questions.
    Mr. Turner. Well, I certainly appreciate the testimony of 
all of the members of our panel, and some of the words that I 
wrote down that each of you used as you were describing this 
process is ``uncertain,'' ``variable,'' ``errors,'' ``limited 
data,'' or ``estimates.''
    Keeping in mind that the wind has always been used as an 
analogy for ever-changing and unpredictable, I know that what 
you are attempting to do is something that is very different 
than what our expectation is.
    Also in understanding the science application of it, it is 
clear that what you are approaching is the theoretical, and 
many people are appearing to look at this information on a 
nontheoretical basis, and real decisions are being made, 
decisions concerning exposure levels, evacuation plans, 
response. It seems that some of these decisions are certain and 
conclusive, but in listening to your testimony today, it would 
seem to me that each of you agree--and that is going to be my 
question to you--it seems that each of you agree that making 
any certain and conclusive decisions based upon this data would 
be incorrect; that the processes are scientific, they are 
improving, and they are certainly important to our overall 
safety and our planning. But we are currently looking at a 
process that may have a margin of error of 100 percent.
    So I would ask if that is true, if my impression is that 
each of you, though committed to the process and its 
importance, would also agree that certain and conclusive 
decisions should not be made based upon any of the current 
modeling outputs.
    Mr. Rhodes.
    Mr. Rhodes. I don't know that I would go that far. If I am 
thinking about urban evacuation, for example, you have to 
evacuate, and in the process of modeling, the application of 
the model to understand the best probability of escape route to 
move people away from the dispersion, that may be the only tool 
you have until you have extremely good chemical detectors 
deployed throughout an urban area or something like that.
    Mr. Turner. Excuse me, Mr. Rhodes. Saying it is the only 
tool you have is different than saying that it is going to be 
accurate.
    Mr. Rhodes. Yes, that is true.
    Mr. Turner. I understand that we may not have anything 
else, I understand the importance of it, but it does appear to 
me that each of you are saying, as each of you review each 
other's data and other types of processes, that drawing any 
real certain and conclusive decisions as a result of modeling 
is currently not advisable.
    Mr. Rhodes. Real certainty. ``Certain'' is the operative 
term there. Certain decisions, I would say, have to be couched 
in understanding that you are making--you are using a model to 
establish probability, and, therefore, the certainty of what 
you are doing, as I said, if it is the only tool available to 
you, then you may have to accept your probability, but it is 
not going to be perfect.
    Mr. Turner. Dr. Johnson-Winegar.
    Dr. Johnson-Winegar. Thank you.
    I would certainly like to characterize it as what I view as 
a continuum across the certainty to noncertainty. And as more 
data becomes available and the models become more robust, you 
can certainly put more confidence in the output of that data, 
which can then be used to make these kinds of decisions. I 
would certainly like to envision it in sort of a phased 
approach or perhaps a tiered approach in that perhaps an 
immediate decisionmaking process is going to be less certain, 
but that as more information becomes available, for example, 
more data points on either the meteorological conditions or 
more information that is known about the particular source. And 
while that might not be available in, say, for example, the 
first 15 minutes, it may be available in a matter of a few 
hours. And again, as was pointed out earlier, specifically in 
the case of biological agents, some of the epidemiological and 
symptomatic data may not be available, as a matter of fact, for 
several days.
    So what we may have to do is an iterative process, where we 
run the first model, we use what data we get from that, what 
output, to make some presumptive decisions. We do it again at 
some other point, whether that is hours or minutes later; it 
depends on how many sources we have for data coming in, how 
quickly that can be analyzed.
    Please bear in mind that while I am certainly not the 
subject matter expert and would defer to many of the others 
here, many of these models are indeed very complex and may 
take, as a matter of fact, hours at very large supercomputers 
to be able to do the generation. So we may indeed have to refer 
to what I like to characterize as a phased approach to using 
the modeling data to help us make those kinds of decisions.
    Dr. Ermak. I believe that the uncertainty is very much 
correlated with the amount of information that we have or the 
data in order to do our simulations. When there is very little 
data, the uncertainty becomes high, and when there is 
considerable data, we can then bring that uncertainty down to a 
bounds that we find acceptable.
    For the purposes of emerging response, I think there is 
also another set of data that we have not talked much about, 
and that is sensor data of the agent or hazardous material that 
has been released. Today there is considerable effort going on 
to develop and to disseminate sensors for chemical, biological 
and nuclear material, or nuclear radiation. The use of this 
data can be used to help reduce the uncertainty in real-time 
responses.
    Our experience in many different types of events has been 
that initially when an event occurs, the uncertainty is quite 
large. While we might have access to the real-time 
meteorological data, we know very little about the source, but 
we are able to predict the pathway in which the cloud may be 
going. From this, first responders can go out, make 
measurements or collect data that would help to verify the 
initial plume and also help us to quantify how much material is 
being released. We find these latter stages are an iterative 
process in which as more data becomes available, we are able to 
make more and more certain calculations of the dispersing 
plume.
    Mr. Hicks. From my perspective, the key word here is 
``probabilistic.'' All the models produce answers that are, in 
fact, statistical in their very nature, so they are 
probabilistic answers. The trick, I think, is that we have to 
learn how to predict the boundary of an area, defining where 10 
percent of the population at least will receive a dangerous 
dose. In this application, I am not quite sure how that would 
be applied. However, I do concur with what was said here to my 
right.
    Dr. Barron. Well, I think there will always be a level of 
uncertainty, but I like to think about what the future might be 
like. I suspect that we will get to the point where we will see 
a distribution of instruments, say, within an urban environment 
that is sufficiently detailed to characterize the main features 
of the flow through that particular city. And then if you can 
imagine an operational mode of forecasting that goes along with 
that process for day in and day out, you are learning from 
the--applying the discipline of forecasting to that region day 
after day, or combining it with experiments and test cases, I 
think that what you will discover is that not only will we be 
able to do a much better job, but you will be able to 
communicate the level of uncertainty. Often it is not a matter 
of whether or not you can eliminate completely that 
uncertainty, but if you have an understanding of the level of 
uncertainty, then you can make sure you don't put people in 
harm's way or you have a much better estimate.
    So I really see this as sort of a transition between 
research, which has been the history of much of this problem, 
moving into this operational phase for which you bring this 
discipline of forecasting to this mode, to this mode of 
operation day in and day out, to the point where you become a 
service, which means the stakeholders are at the table, and you 
have learned from each other, atmospheric scientists from that 
community of responders and responders in terms of what the 
atmospheric science community can deliver, and you can give a 
good estimate of what that level of uncertainty is. Then I 
think you have accomplished a great deal.
    Dr. Hanna. Concerning our confidence in the models, I would 
like to point out that the EPA uses just about these same types 
of models thousands of times over the past 20 or 30 years to 
make decisions about emissions controls for plants, which has 
then been followed up by observations about the plants, and 
these models have been shown to be reasonable for those 
thousands of applications, which is similar to this 
application.
    Concerning acceptance criteria and looking at all of the 
various comparisons with observations, it seems like once a 
model gets within 30 or 40 percent of the observation, that is 
what can be considered an excellent substance criteria, but 
that is if you have a lot of onsite data. Once you get to a 
situation like Khamisiyah with hardly any meteorological data 
or anything else, I suppose it degrades to a factor of 5 or so. 
But we do have a lot of evidence of model accuracy that is 
built up over the years.
    Mr. Turner. I appreciate your answers. The reason why I ask 
the question is we have had testimony in front of us that 
relates to the issue of what happens when these models that you 
are working with get in the hands of others, because we have 
had individuals who have testified that we can predict the 
plume as a result of a specific incident, and each of you being 
experts in the field are saying that of course they are useful, 
they certainly are better than anything else that we have, they 
give us information that is necessary to determining how to 
react, but yet they are not specifically conclusive and should 
not be absolutely relied upon. I wanted to hear your responses, 
as I know that we have heard in other hearings individuals 
talking about the absolute prediction of plume incidence.
    One other question that came out of Mr. Hicks' testimony. 
You have in your written statement, ``The coarser product is 
routinely made available via the Internet to users who are 
registered by scientists of my laboratory. This is the Realtime 
Environmental Applications and Display System, used routinely 
by over 1,500 registered users for accessing and displaying 
meteorological data and running trajectory and dispersion 
models on the Web server of my laboratory.''
    One of the things that we have been hearing about also in 
this committee is issues of tracking data and the types of 
access that individuals have to data. So one of the things that 
I would like to know both from you specifically, for example, 
what types of reviews do you have of who is having access to 
this information, and what they are using it for, but also from 
the other panelists as to this information that you receive 
gets specific enough that your models are able to predict with 
accuracy, to what extent do we need to be concerned about 
having a classified nature to the outcomes of your work?
    Mr. Hicks.
    Mr. Hicks. Yes. Immediately after the September 11 
incident, we closed down access to the Web operation except to 
users who were either from a dot mil origin or from a NOAA 
origin. We then opened it up to registered people. In other 
words, we went through the process of checking out the 
credentials on people as they came in. Only the coarsest data 
are available that way. The 40-kilometer Web data are used, and 
they are made available. The fine-scale stuff, the fine-scale 
data that are necessary, for example, for predicting what might 
happen in Washington, DC, New York City and so on, those data 
are not then made available through that source.
    Mr. Turner. Do you track also then what people are doing 
with the data that you do provide to them? Are you aware of 
what--because it said access to your server. You do know what 
people are doing with the information you are providing?
    Mr. Hicks. Yes. We keep track of the runs that are made, 
and we make sure that we know exactly who is using them for 
what.
    Mr. Turner. Other members of the panel, any concerns that 
you might have about the information being available to 
individuals that might use it to cause more harm than good?
    Mr. Rhodes. Well, it is a genuine concern. It is on the 
same scale as imagery. If you are going to Space Imaging, and 
you get a photograph from outer space, it is at a certain 
granularity. One of the conundrums associated with it, however, 
is that this is math. There are lots of people on the planet 
who can do the math without having to come to NOAA. So even 
though they may not have access to the fine-grain information, 
you raise a legitimate concern about how much information do 
you want to disperse to whom, because then the tool is now 
turned as a tool for your opponent.
    Mr. Turner. Anyone else want to speak on that issue?
    If not, Mr. Chairman.
    Mr. Shays. Thank you. It would strike me that it is much 
more difficult to predict where a plume has gone than it is 
later to reconstruct it and say where it has been; is that 
accurate or not? Mr. Rhodes.
    Mr. Rhodes. I hate to sound like a broken record, but it 
depends on where you had the information. If you can 
reconstruct the information, reconstruct source term, 
meteorological data from the time of the event, then after the 
fact it will be easier to reconstruct if you didn't have that 
information at the time of the event or a priori.
    One of the concerns about the Khamisiyah event in and of 
itself is that the data are limited, and, as you know, Iraq 
quit submitting meteorological data to the World Meteorological 
Organization in I believe it was 1981. So no one was able to 
collect meteorological data except from sites that were distant 
from the Khamisiyah site. So unless you can get that detailed 
data up front, or after the fact, then the reconstruction is 
difficult.
    Mr. Shays. Anyone choose to add to that?
    Dr. Barron. If you have good knowledge of the source term 
and good meteorological observations, then in hindsight you 
have the advantage that you have the time to run multiple 
realizations of models, and, therefore, you can have an 
ensemble that gives you a better sense of the probability of 
the distribution, if you have the data to work with. Whereas if 
you were looking actually during an event, and you were working 
to respond quickly, you might not have the time for multiple 
realizations.
    Mr. Shays. OK. I am going to use Dr. Johnson-Winegar's 
explanation on page 2 of her statement when she is talking 
about the variables. She said basically, weather conditions are 
an obvious factor, such as temperature, wind speed, cloud cover 
and so on; geographic conditions, such as topography 
structures, type of vegetation, type of chemical, biological, 
or biological threat agent; and the state of the threat agent. 
Is that all one part, Dr. Johnson-Winegar, or is that two? It 
said the type of chemical or biological threat agent and the 
state of agent, the fifth?
    Dr. Johnson-Winegar. Yes. They are separate.
    Mr. Shays. OK. And then the type of delivery systems and 
the type of event.
    Would you add anything to that as I went through it, which 
is--let me just deal with No. 3, which is the type of chemical 
or biological threat agent. Is a chemical or a biological 
harder to model, or is there no difference?
    Dr. Johnson-Winegar. I will start out, and my esteemed 
colleagues can chime in. I think it gets back to the issue that 
we have made repeatedly in today's discussion, the source data, 
and so currently I would assess the fact that our overall state 
of knowledge about chemical weapons is more defined and more 
well understood than the biological agents, so that is just one 
piece of the information.
    Mr. Shays. How about radioactive, radiological?
    Dr. Johnson-Winegar. I think that is even better than 
chemical. We know more about that, and I would put that in a 
more advanced stage than chemical. And I am speaking primarily 
of what are known as the traditional chemical warfare agents. 
If, as a matter of fact, you would want to expand that 
definition to everything including toxic industrial chemicals 
and toxic industrial materials that may be a greater concern 
for a civilian incident than a military incident, then 
obviously our total body of knowledge goes down somewhat.
    But with regard to the biological agents in particular, 
that is where I assess that we have some of the largest data 
gaps in knowing things about the various types of agents and, 
in particular, as I mentioned later on in my statement, the 
actual effects on humans via aerosol exposure of many of these 
biological agents. We have limited ability to extrapolate from 
animal studies and certainly in many cases no human effects 
data of many of the biological agents. So that brings us back 
to the point that all members of the panel have made, without 
being assured of a lot of the source data, then that has an 
impact on the output from the model, to be sure.
    Mr. Shays. Does anyone else on the panel want to speak to 
the issue of chemical, biological and radiological?
    Dr. Hanna. In my statement I mention the difference between 
biological and chemical in that you don't know that there was a 
biological release in general, so you can't really do an 
emergency response calculation, and there is some research 
centers around the country that think that atmospheric modeling 
is not of much use to biological incidents.
    Mr. Shays. I sense that. But what about radiological or 
chemical in general?
    Dr. Hanna. Well, chemical you know that there was a 
release, and radiological you also tend to know that there was 
a release.
    Mr. Shays. But do they respond basically the same way?
    Dr. Hanna. Yes. They would transport and disperse the same 
way.
    Mr. Shays. What I am trying to understand is--in the end is 
that if we know the weather conditions--I mean, we have a sense 
of the topography, but if we know--and the type of vegetation, 
those are fairly obvious. We can make some quick assumptions 
about that. But weather is obviously going to be one big 
variable. Is forecasting the weather a science or an art? And 
it is a science, but its impact in the end is an art, from my 
standpoint.
    What I am just trying to understand is where are the big 
challenges in those six key types of information: the weather, 
geographic condition, what type of agent, you know, chemical or 
biological, the issue of the state of the agent, the type of 
the delivery system and the type of event? It just strikes me 
that weather is--I have always assumed that weather was the 
biggest element, and in the end, obviously, the concentration 
of the material in terms of its impact on the populace is going 
to be obviously not just the weather. I understand that part of 
it.
    But what I am really wrestling with right now is--help me 
out here. What becomes the biggest challenge to people in your 
field? Is weather the key?
    Dr. Hanna. Well, I think people that do comprehensive 
modeling with emissions, then transfer and dispersion, and then 
risk assessment believe that the emissions and the risk 
assessment are the largest challenges, that probably the 
weather is--of those three is what we know the best. However, 
when you are worried about wind direction, as in the Khamisiyah 
example, I think we have--we really need to know the wind 
direction in order to do the troop assessments that you are 
talking about.
    Mr. Shays. I mean, we have done the tabletop. Once we knew 
what we were dealing with, a chemical, and what type of 
chemical, the key thing was--we asked--we wanted to know which 
way the wind was blowing and how fast the wind was blowing and 
what was the humidity. I gather that has something to do with 
its--what, what would humidity tell us? So it is mostly wind.
    Dr. Barron. Well, humidity could affect a particular agent, 
like a mustard agent or a nerve agent; it would affect the 
deposition if there was rainfall, the deposition of the agent 
out of the atmosphere. That is the reason why having that 
humidity and precipitation elements are valuable.
    Mr. Shays. What I am surprised about is I am not seeing a 
lot of people jump in here. Why are you not trying to help me 
out?
    Dr. Barron. Well, I think there are a lot of uncertainties. 
My view is that urban meteorology or anyplace with complex 
topography and an observational suite which is less dense than 
the scale of the circulation that would be going through 
buildings, and very little practice at obtaining this 
discipline of forecasting within that region, that is a 
substantial problem. We have little experience, and this is a 
vulnerable region of the country. So I believe that is a 
substantial challenge, along with the other elements.
    Mr. Hicks. If I may come in, the first two you mentioned, 
the weather and the geography, the geography is important. The 
topography is important because it affects the wind direction 
as much as anything else. So that is tied in intimately with 
the weather, the meteorology of the problem.
    Our perspective at the moment is that the key thing we have 
to worry about is to make sure we can do what we say we can do 
in the areas where people actually live, where people will be 
affected, and that is, we are finding, a very, very difficult 
thing to do. The urban areas are difficult to address, because 
the buildings do interfere with the wind so much. I think in 
Dr. Johnson-Winegar's language that would be topography.
    Mr. Shays. Anybody else?
    Dr. Ermak. I would say that I think weather is an important 
uncertainty, and especially in the urban areas, both because of 
the complexities of dealing with the flows around urban areas 
and because that is where our populations are located. I think 
we need both additional data such as the Washington DCNet that 
was being set up in other urban areas to support our work 
there, and I think we also need research into urban dispersion 
modeling and understanding the flows in these areas.
    Other areas, I think, that create uncertainties have to do 
with the fact that many of these agents are not in a gaseous 
form, but either in an aerosol or a particulate, and 
understanding these could have a dramatic impact on how far 
they disperse downwind and where they settle onto the ground. 
So that is another.
    And I think particularly with biological, a third thing 
that must be--that is not well known is viability. Because the 
agent is in the atmosphere and it travels downwind doesn't mean 
that a person who was exposed to it and is still alive that it 
could cause illness and other difficulties to the person. So I 
think that is another area where research is needed.
    Mr. Rhodes. I would say I guess there is a variation on 
weather, but understanding the time, the duration of the event, 
if something occurs at sunset and extends into the evening, if 
something occurs in the middle of the night, if something 
occurs at dawn, these are factors that have to be worked in, 
because now you have temperature layers that are different. 
Talking about a source term and talking about saying--making 
the statement ``I understand the chemical'' is an extremely 
broad statement, because that means--for example, in Khamisiyah 
that meant you understood how many rockets, of what type, in 
what container, in what configuration; and they were blown up 
with what; how much was ejected; was it in a pit, was it in a 
bunker; was it at night, was it during the day.
    So there is an awful lot of data that, when we are talking 
about the data that we need, the source term data, there is a 
tremendous amount of data that we need. If this is an 
evaporative chemical; is it a persistent chemical? As you heard 
2 weeks ago in our discussion about Wallingford, anthrax at 
less than a 5-micron diameter operates as a vapor, and as you 
saw there were 3 million spores underneath the No. 10 machine, 
and yet we found spores 25 feet above it in the high bay. So 
does it settle? And when it settles, is it stable? I mean, all 
of those factors are involved. But the time of day when 
something occurs is extremely important, because then you 
understand what the varying temperatures are between the 
ground, which may still be warm, and the air that is cooled in 
the desert, for example.
    Mr. Shays. You could make an argument if the plume is like 
the Livermore plume, it seems to be broader. You could 
potentially make an argument, it would seem to me, based on 
science ultimately, that though more people were exposed, the 
concentration may be so much less that the exposure isn't 
serious; whereas if it happens to be the more concentrated 
plume, that it is likely--but obviously, then, we want to know 
what was blown up at Khamisiyah. I understand all of those 
factors.
    Mr. Rhodes. I guess one of the following points leveraging 
off of what Dr. Johnson-Winegar said, looking at Dr. Hailey's 
work down in Texas, trying to establish what is minimal 
exposure, what symptoms, what conditions are expressed over 
time based on what exposure, that is a key item, so that even 
though you are talking about the olive green and the yellow 
area, those people may just express symptoms later, like ALS or 
something like that.
    Mr. Shays. Would you all react to this, because I am trying 
to sort this out. Based on a number of hearings our committee 
has had, tell me if I am on the right track or not. Obviously, 
for our veterans, we want to reconstruct where the plumes went, 
and we want to know the impact of the plumes, the concentration 
of exposure and so on. But in our fighting this war on 
terrorism, the bigger need is to be able--at the moment of an 
attack to be able to have a sense of who is potentially in 
danger and who isn't, and where you are safe and where you are 
not safe. And so it seems to me that what we are really 
trying--and maybe we gain from reconstructing the past. I mean, 
we do, but it seems to me our primary efforts should be--and 
this committee's primary effort and the government's primary 
effort should be on how can we have more accurate projection of 
plumes when there is an attack so that we are sending people to 
safety and we are treating the people who may need to be 
treated. Is that a fair statement? Does anybody disagree with 
that?
    The reporter cannot take a nod and a shake. So the bottom 
line, I am seeing a lot of heads go up and down. Anyone want to 
say it better for the sake of me and the reporter?
    Dr. Ermak. Let me come in for a moment. I agree with you 
completely. What we most recognize, I feel, is that we do have 
a lot of meteorological information available. The models that 
are available today are making good use of part of that 
information. We have to get to the point where we can mine the 
total information body, the total network of information.
    Mr. Shays. So we can instantly say that most of the time 
the wind direction goes this way most of the time, or when the 
temperature is this, and this time of year, so we could almost 
turn to the computer and make an assumption. If we didn't have, 
you know, some accurate, present information, we would just go 
historically and make assumptions; is that what you are saying 
to me? If you could tie this in, if you would, with your issue 
of the urban sensors.
    Mr. Hicks. Yes. What I am trying to say is that in the 
final product, every emergency manager would have the ability, 
would have the information in front of him that would draw not 
only upon the best weather forecast information available, but 
also upon those data sets that are within his own area, and 
they may be the Department of Transportation's data, they may 
be the Environmental Protection Agency's data, they may be data 
from private sources. These data have to be, to my mind, 
exercised. We have to learn how to make use of all of the data 
sets that are available out there, because a lot of data are 
available in urban areas that are not being used at this time.
    Mr. Shays. Anybody want to add to that?
    Dr. Johnson-Winegar. What I would like to add to that is 
with regard to your comment on what the government can and 
should do to increase our predictive capabilities, I certainly 
think that is a very important aspect. Some of the things that 
the Department of Defense has been doing and is investing in 
for the future includes such things as improving the 
sensitivity and specificity of the various types of sensors 
that can be deployed either for military use or for civilian 
use, and that goes to the things that are being used in 
BioWatch and a number of other different scenarios.
    Also we talked about low-level effects. We have embarked 
upon a very ambitious program to look at low-level, i.e., 
subacute, impacts of a number of the different chemical agents 
known to us.
    Third, I would like to point out our program in what is 
called agent fate and the fact that there are, again, a number 
of assumptions that are used as to whether agents would be 
absorbed into various surfaces, concrete, sand, whatever, and 
what is the possibility of either reaerosolization in the case 
of anthrax spores, for example, or off-gassing when the 
climatic conditions may change or something like that. These 
again just point to a lot of the unknowns.
    I am sorry that the whole panel keeps coming back to that 
point, but it is a very important point to make to you, that 
with regard to what we know about the source term data, what is 
the agent going to be? What kind of form is it going to be? How 
is it going to be impacted by the meteorological, as was 
mentioned earlier, in high humidity or in rain, or, as was also 
mentioned, the time of day, the inversion layers in the air? 
All of these things have to be fed into the model. These are 
all areas that are crying for an additional investment in the 
research programs, and I think that we can be proud of the 
investment the Department of Defense is making in some of those 
areas.
    Mr. Shays. Let me be clear and just make one last point 
before I am all set with my questions.
    The Department of Defense, though, is not--its focus is not 
on terrorist attacks in urban areas in the United States; is 
that correct? I mean, whatever work you are doing, you are 
doing more on the battlefield than you are in an urban setting, 
correct?
    Dr. Johnson-Winegar. Well, our primary emphasis is on the 
battlefield, but I think in many areas, the information is 
easily transferable to an urban setting. For example, agent 
fate, you know, is it going to be absorbed into the concrete on 
the street, probably the same type of data would be generated 
as to would it be absorbed into a runway on an airfield.
    Mr. Shays. This is my last point. When we had our hearing 
on the issue of anthrax as it related to our government 
buildings here and in the post office in Wallingford, they 
did--twice they tested in the facility. They did not believe it 
was where they tested. When Ottilie Lundgren died, they went 
back and they found it. And one of the points that was made to 
us was that had they known it was there, they would have found 
it, but it was so small. I mean, what they were looking for is 
such a small--it is difficult to find it, and the bottom line 
was had they known it was there, we would have been we know it 
is here, now we have to find it, as opposed to it is probably 
not here, and they did the test, and they didn't find it. When 
they knew it was there, they did find it. That is my point.
    We know that there will be chemical, biological, heaven 
forbid maybe even a nuclear attack on the United States. We 
know that a prime target is a place like New York City. So if 
we knew--like right now you just knew that sometime soon there 
would be an attack, a chemical attack, say, on Times Square, 
would it take us--do you agree that we would probably very 
quickly, in preparation, be able to prepare for it, know the 
way the wind basically goes, and be able to say not what is 
going to happen 10 or 20 miles down, but what is going to 
happen 15 blocks away?
    My point is if you knew it was going to happen, do you 
think that we would be making a lot faster progress? Do you 
understand the question?
    Dr. Ermak. Allow me to answer that. If you knew or 
thought--perhaps a better way to put it is there was a high 
probability that something might happen, or you had information 
that it might, you can, of course, do a better job of preparing 
for it. At NARAC, we have been involved in situations such as 
that where at certain places certain events were occurring, and 
it was anticipated that this might happen.
    Now, in all of the events that we supported, there was not 
a release. However, we were able to bring much more of our 
resources to bear on to that situation, both in the collection 
of data and in the running of more high-fidelity models to 
address it. So I would think--I would say at least yes; the 
answer is if you knew that an event might be occurring, you 
could be much better prepared in having plume predictions with 
greater fidelity and accuracy.
    Mr. Shays. I am taught to observe, and one of the things I 
am observing is that you all do work that no one knows much 
about or really cares that much about, and they should. And I 
had this sense that, you know, you just kind of plod through 
this, you have done it for years, and you keep doing it. I 
guess I would like there to be a higher sense of urgency. I 
would like to feel a higher intensity level. I would like to 
feel like--you know, someday you all are going to be on TV 
having to respond to some attack somewhere, and they are going 
to ask you about this boring thing called the plume, and you 
are going to try to explain it to people, and then you are 
going to think when you go home, my God, if we just did a 
little more a little bit sooner, it might have helped. That is 
kind of my sense of what I am gaining from this hearing.
    I am set to relinquish my time unless someone wants to make 
a comment.
    Dr. Hanna. I think I would like to second what Bruce Hicks 
said. There is no substitute for wind observations in the urban 
area, and there are a number of them that are already in and 
being proposed for the Washington area, for the New York area, 
and there is the urban atmospheric observatory being proposed 
for New York right down in the Times Square area. And you just 
have to have those local wind observations to tell you which 
way the plume is going to go, because you don't want to use the 
Baltimore airport wind or the LaGuardia wind or something.
    Dr. Barron. I would just like to add that the Board on 
Atmospheric Sciences and Climate report wasn't one that was 
requested, it was one that the atmospheric sciences community 
felt that this was essential to begin to take these steps; for 
instance, instrumentation and modeling of a city to work on the 
forecasting, gain experience, do model intercomparisons so you 
would know which model was accomplished in what particular 
facet. So that was entirely our intent was to provide a path 
for how atmospheric sciences could best address the issue that 
you raised.
    Mr. Hicks. And I would like to volunteer that neither the 
DCNet in Washington, DC, or the Urban Atmosphere Observatory up 
in New York were generated top-down. They were both generated 
by the scientists recognizing there is a real problem here, and 
we had better start addressing it fast or else we will get into 
trouble. We are trying to dig our way out of a hole, and do it 
fast.
    Mr. Shays. Thank you, Mr. Chairman.
    Mr. Turner. I don't have any other questions, but we do 
have some questions from our counsel.
    Mr. Halloran. Thank you. I just want to first ask for a 
couple things for the record from Dr. Winegar, if I could.
    Your statement describes the J-1 process as semi-automated. 
I wonder if we have a more complete explanation of what is 
automated and what is not, in some more detail, on what the 
plans are to automate that system.
    Dr. Johnson-Winegar. I will be happy to take that for the 
record and provide you more details of the various phases of 
the full integration of J-1.
    Mr. Halloran. Thank you. Your statement early on says that, 
until recent reorganizations, you were the party responsible to 
accredit models. Who does it now?
    Dr. Johnson-Winegar. Under the recent reorganization, my 
immediate boss, Dr. Dale Klein is now the accreditation 
authority, because his purview reaches across nuclear, as well 
as chemical and biological. So I believe that is the 
appropriate person.
    Mr. Halloran. Thank you.
    Mr. Rhodes, briefly, to what do you attribute DOD's 
limiting of the plume height or altitude in their modeling? 
What drove that?
    Mr. Rhodes. I have no answer for that. I believe it was--as 
we understand from the documentation, it was an arbitrarily set 
value, and it was described as an arbitrarily set value.
    Mr. Halloran. I see. Was it in your testimony or someone 
else's that described the videotapes that we had seen here of 
that event looked to show plumes higher than that.
    Mr. Rhodes. It was actually in our document, in our further 
testimony and in our subsequent report, talking about the 
thousand-pound bomb estimate and plume being as high as 400 
meters. That was us.
    Mr. Halloran. Thank you.
    And finally, Dr. Ermak, could you just briefly tell us 
about your experiences in TOPOFF 2 and what the reported 
problems were with the plume applications and modeling in that 
exercise scenario?
    Dr. Ermak. Yes. We participated in TOPOFF 2 in two ways. 
One was in direct support to the city of Seattle and the 
surrounding area, and the other was through the Federal 
Government, through the Department of Homeland Security and the 
Department of Energy. I think one of the points that I think 
came out of that, it is not only important to have the accurate 
data and accurate model predictions, but also important is to 
be able to rapidly provide emergency managers and the first 
responders with that information, and to provide that 
information in a way that they can readily use it.
    As an example, during that exercise we had one of our staff 
in the city of Seattle supporting the Fire and Emergency 
Operations Center people in Seattle in the use of our system 
that was being tested. He was also available to answer 
questions that came from, say, the mayor or other public 
officials. And this was very, very useful. And so that says 
also that in addition to--and he emphasizes the need for 
information that is in a form that they can use to make their 
decisions.
    Mr. Halloran. And what is the impediment to that? I mean, 
it just came in a form that he didn't understand, or no one was 
willing to make a definitive call as to what it really meant?
    Dr. Ermak. No. I think, sometimes for example, just a plume 
picture is not readily understandable, say, by a policymaker or 
decisionmaker. And so putting that into terms in which they can 
understand it, understand the reliability of it is, is very, 
very helpful to them.
    Mr. Halloran. Thank you, Mr. Chairman.
    Mr. Turner. Thank you. I want to thank each of you for 
participating in this. Certainly the work that you are doing is 
important and very complex, and we respect the expertise that 
you bring. I think the issue before the committee has focused 
somewhat on how that information is used. And as our chairman 
has raised the issue of, as you look to modeling, what 
resources are going to be necessary for your success.
    With that, I want to ask if anyone has anything else they 
would like to add to the record? Hearing none, I want to thank 
you again for your participation, and thank our chairman. We 
will be adjourned.
    [Whereupon, at 3:01 p.m., the subcommittee was adjourned.]
    [Additional information submitted for the hearing record 
follows:]

[GRAPHIC] [TIFF OMITTED] T1396.103

[GRAPHIC] [TIFF OMITTED] T1396.104

[GRAPHIC] [TIFF OMITTED] T1396.105

