Thursday, July 17, 1997 - 1:50 p.m. (EDT) Dr. Bernard D. Rostker, Special Assistant for Gulf War Illnesses
[Also participating in this briefing are Mr. Bob Walpole, Special Assistant to the Director of Central Intelligence for Persian Gulf War Illnesses Issues; Dr. Susan H. Mather, Assistant Chief Medical Director, Public Health & Environmental Hazards Office, Veterans' Health Administration, Department of Veterans' Affairs; and Dr. Jeffrey Grotte, Project Leader, Institute for Defense Analyses]
Dr. Rostker: Good afternoon.
This is the last of our information presentations concerning Khamisiyah. Next week we will present to you the results of our analysis. It's coming along fine, and we're quite confident we will be able to do that.
We wanted to bring before you, for your review and consideration, the work we did at Dugway to understand better the source terms, the work we've done to better understand who was at Khamisiyah or in the vicinity of Khamisiyah in those first weeks in March, and now the long-awaited IDA report on the modeling activity.
I have with me today two of my colleagues, Bob Walpole from the CIA. Bob and I have worked jointly with IDA and with the modelers to bring forward the best possible models and representations that we can of what happened in Khamisiyah, and Bob will explain that and also touch on our joint efforts to reduce some of the uncertainty.
The bottom line, of course, in Khamisiyah is who might have received a dosage, who might be ill, and this gets to the issue of the medical aspects, particularly of low level chemical exposure. As you will see next week the vast majority, 99 percent of the area that we will be showing you in a plume, in a laydown, has to do with low level chemical exposures, not chemical exposures at an acute level or a level that would cause death. That's a very small area, and we'll explore all of that next week.
In regard to the health issues and also the issue of health benefits and exposure, we're joined today by Dr. Susan Mather from the VA, and after Bob has made his presentation, Susan will talk to you about this from the VA's perspective.
When we're all finished, we'd like to hold questions, and then we'll all come up and be happy to answer any questions that you have.
With this, Bob?
Mr. Walpole: Thank you, Bernie.
Good afternoon. You should have a package that looks like this that we prepared to walk through some of the uncertainties. If you recall, the whole modeling effort is... Okay, you don't have that yet.
The whole modeling effort is part of a larger, robust effort addressing the Gulf War Illnesses issues. We're pursuing this in conjunction with the Department of Defense to try to help in that regard.
Last year our modeling efforts came to a halt because the uncertainties were so great in the source term, in the meteorological data, that we could proceed no further. The first page of the packet that you have walks through those uncertainties. We had two soldiers that were telling us about events in the pit, their stories were contradictory. It wasn't clear whether it happened on one day or multiple days. We were somewhat uncertain about the source term, and greatly uncertain about meteorological data. There had been no testing of what happened to a chemical agent in an open pit demolition to give us a handle on how the agent would aerosolize in that environment.
The second page walks through what we have done the last several months to address and reduce those uncertainties. First, we have increased our interviews with the soldiers and increased the number of soldiers that we're talking with. We're now talking with five soldiers, and that appears to be at least half of those that were involved in the pit.
They have eliminated uncertainty on the date of the pit. They indicated that the March 12th entry in a log was recorded after the fact from accumulated data and that we shouldn't put much credibility in it. They recall doing activities in all areas of the pit. So we're confident now that we have the one event on March 10th.
We have the IDA review paper that I'll mention in a moment here. We have done a lot to reduce the uncertainty in the source term. As you see in the paper here, we now have a best estimate of 1,250 rockets in the pit at that time. We have an amount of agent per rocket at 6.3 kilograms.
This is interesting, because we didn't discover this until we started into the Dugway testing. As we manufactured warheads for use in this testing, we determined that our estimate of eight kilograms included the mass of the plastic container for the agent, and that it was really only 6.3 kilograms.
The Dugway testing provided other information as well, such as how much aerosolized and how much spilled into the wood and into the soil.
We also discovered new information on meteorological data. This ranged from looking at soot plumes that were created from bunkers that were destroyed the same day as this pit demolition so that we could see what direction the winds were blowing, as well as tracking the smoke from the oil fire plumes. And finally, we have used multiple mathematical models and modelers in the effort. In the effort we had last year we had one model, one modeler and two models. IDA had recommended that we should link the models. We had done a linkage last year, but IDA did not feel that was sufficient. Now with multiple models and modelers involved, we have multiple linkages, that that appears to be addressing the concern.
A few pages into your package, most of the rest of the slides are all background information, so I'm going to skip about three pages into your package where I have the IDA panel review -- reiterated the uncertainties that I have just stated, and recommended several things that we should do.
One, IDA recommended that we should examine multiple parametric variations to bound the source term. IDA made this recommendation before we had done the Dugway testing. The Dugway testing and a lot of new information that we have been able to uncover have reduced the source term uncertainty enough that we don't need to run multiple variations, but rather have a best estimate on what was there at the time. IDA recommended that we run the meteorological models for at least 72 hours, and some of the models we have been running this week have been going 100 hours, so we are definitely meeting that concern. Finally, IDA recommended that we link the models, and I discussed that a moment ago.
If you look at the last page of your package, there are two columns on that slide. The left column just shows you the models that have been involved in this activity -- two types of weather models, and then the transport and diffusion models at the bottom. Over on the right hand side are those agencies or components that are involved in the modeling effort. As I said before, one of the ways we're addressing uncertainty in this issue is to have multiple models and modelers involved.
That concludes my remarks. Dr. Mather, would you like to follow?
Dr. Mather: Thank you. I'm just here to represent the VA, and to thank DoD and CIA for their efforts to better understand the potentially hazardous exposures that Persian Gulf veterans have experienced in the Persian Gulf. We at VA are always glad to have additional information on these exposures, but I think we have to keep in mind that our basic goal is to provide quality health care benefits and services to Persian Gulf veterans, and that includes keeping them informed about investigations and research efforts of the Administration.
Just to refresh your mind on health care services, I'm sure you're all aware that any Persian Gulf veteran who is ill and feels that his illness or his symptoms are related to service in the Persian Gulf can come to the VA for health care. They can come in and get a free comprehensive physical examination and treatment for any symptoms or illnesses that they have, they will be put on the Persian Gulf Registry, and will then be kept informed through the Persian Gulf Review, and we have a copy of the most recent Persian Gulf Review available for you today.
The VA Gulf War Veterans programs are a comprehensive, four-pronged approach including medical care, research, compensation, and finally outreach and education. VA provides Gulf War Registry health examinations and follow-up care in all its medical facilities nationwide, and there are specialized examinations and facilities available at four regional medical facilities. There's also readjustment counseling available in the vet centers, as well as sexual trauma counseling for women veterans who experienced sexual trauma while they were in the Gulf.
To date more than 66,000 Gulf War veterans have completed the Registry examination, and this is in addition to the 32,000 who completed DoD's comprehensive clinical evaluation examinations. We've had more than 1.8 million visits in our ambulatory care facilities from approximately 200,000 Persian Gulf veterans, and more than 20,000 Persian Gulf veterans have been hospitalized in VA facilities. Nearly 400 Persian Gulf veterans have reported to our referral centers for continuing evaluation, and we've also provided counseling to more than 75,000 Gulf War veterans at VA Vet Centers. If a Persian Gulf veterans is disabled from the experience in the Persian Gulf, VA does provide disability benefits. We have a law in place that allows us to provide benefits even if we are unable to diagnose an illness responsible for the disability. This will be true as long as the symptoms manifest themselves within ten years of their service in the Persian Gulf. So we're well within that period right now.
We're currently providing compensation to 29,000 Gulf War veterans and 1,100 veterans are receiving compensation for undiagnosed illnesses.
To move into the area that we're interested in today, we are trying very hard to get a better assessment of the Persian Gulf experience. One of the things we're doing in that is a lot of research under the Veterans Coordinating Board. The Coordinating Board is made up of representatives from DoD, VA and Health and Human Services. The goal is to have a fully structured research protocol to address all the issues that affect Persian Gulf veterans. More than 90 research projects are currently in progress, and others have been completed. We're going to continue to search for answers to the complex problems that face the veterans who served in the Persian Gulf. Research related to possible low level exposures to chemical weapons has intensified in the wake of the information about the release of nerve agents at Khamisiyah in March 1991, but I think we have to look at the issue of exposures in the context of all the hazardous exposures that Persian Gulf veterans were in contact with in the Gulf.
Because we're not dealing with just one cause or one disease, our epidemiologic research is extremely critical. VA is nearing the completion of what will be the first general survey of Persian Gulf veterans, and that will include 15,000 Gulf War veterans and 15,000 veterans who were serving in the military at the same time, but who were not deployed to the Gulf. This will be very important in helping us to understand the health status of this entire group.
It's important to note that knowledge of specific exposures is useful, but it's not necessary in diagnosing and treating Persian Gulf veterans who are having problems as they have returned home from the Gulf. That is basically our goal. Our focus in VA and I think in the federal government will remain treating individual veterans for his or her specific health concerns, and that we will continue to do.
Dr. Rostker: Bob, why don't you join me up here, and we'll be happy to take any questions, and Susan also.
Q: The last time you met with us, you gave us an estimate of the number of troops within a 50 kilometer radius of Khamisiyah. Now that you're a week away from releasing some kind of estimate of troop exposure, can you tell us if there is any evidence that troops outside that circle beyond 50 kilometers might have been exposed to anything from Khamisiyah?
Dr. Rostker: We do not have a troop count at this time. What we have been working on is the plume and getting that straight. The last ones were last night. We expect that we will have that laydown today, and we will then match that with the troops in the area. But I would tell you that the area is shaped differently than just 50 kilometers around Khamisiyah, and over the 72 hours or in some cases the runs are longer than that, we are talking about an area that is not characterized by a circle and not characterized by 50 kilometers.
Q: Does any part of that fall outside that circle?
Dr. Rostker: Yes.
Q: Now that you have this, maybe it's a question for Mr. Walpole, but we have this data now that indicates, according to this, that 500 rockets were affected by the demolition. You say here that many of the rockets weren't harmed, 60 percent of them in your test. We know that the purity is less than you thought, or the agent purity is 50 percent less agent in each rocket. Do all of these things point to a smaller potential exposure than you might have originally feared from these incidents? When we look at all this data you've provided us today.
Dr. Rostker: I think that's a safe conclusion to draw. Obviously, at one point we were at 100 percent purity; at one point we were vaporizing instantaneously all of the rockets. What we've been able to do is gain a lot of insight into the events at Khamisiyah, and we've applied that insight into the analysis. The analysis is also much advanced by the help we got from IDA in terms of how to put the various models together, as well as some dedicated efforts to find classified and unclassified meteorological patterns.
We went even further than that. Since Dugway there have been substantial tests on the issue of evaporation from both soil and wood, and we have a much better handle on the phenomena of evaporation and how that affected the leaching of these elements into the atmosphere over that time.
Mr. Walpole: I just want to add briefly to that, whenever we don't have a data point, we're going to choose the worst case assessment, on the agent purity, for example. We used 100 percent before because we just didn't have enough information to push it lower, even though we knew that they didn't have 100 percent purity. We didn't have any data to quantify that. Now we do, and so we're able to lock it in at 50 percent and feel confident that we've got that. That's true for other indicators that we've now listed there. But any place where we don't have a data point, we're going to go with the worst case side of that.
Q: Can you just comment on this published report in the Army Times this week that suggested that there's some sort of internal memo in the Pentagon warning that up to 80,000 U.S. troops could have been exposed?
Dr. Rostker: I can tell you I know of no memo in the Pentagon that had any number at all, to say nothing about a number that had 80,000. We have been concentrating intensely on the issue of what does this plume look like, and the number of troops will fall out from that. I just don't know what that number will be, and it hasn't guided any of our considerations.
We have prepared two letters -- one for those who we previously corresponded with who we can, with a great deal of certainty, say were not exposed to even low levels of chem; and then a letter for those who we believe may have been exposed to low levels of chemical and reiterate the now important common statement that if you are not feeling well it is important to go and join the health registry. Those are the two populations we are dealing with, but I don't know and I probably won't know until early next week what the actual numbers count are. We have a moving cloud and we have moving troops under that cloud, and they have to be correlated to determine what the exposure levels were, and we'll do that on a day-by-day basis over approximately a three day period. And the three days is not arbitrary. That is the maximum extent of the spread of this cloud. We gave it plenty of opportunity to spread further.
I might also say that we have multiple models, and what we have done is take the laydown of those models and draw a footprint around all of them. So we've captured any projections that are in the multiple models that we're using, and that's what we'll present next week.
Q: Bernie, if I could ask Bob... On the Dugway testing, Bob, it's a little scientific inquiry here. The simulant that's used in Dugway, how close to an imitation of sarin is it in its physical properties -- its dispersion, its evaporative properties? Let's talk a little also about the moving cloud. What do you, from Dugway, expect was in that cloud? What was coming down to the desert floor out of that cloud? And one more after that, but go ahead.
Mr. Walpole: The simulant that was chosen was chosen specifically because it replicated or closely matched the evaporative volatility characteristics of Sarin. That's what we wanted to understand, was the aerosolization.
The test at Dugway, as you can see in the numbers here, we had about two percent of the total, and I'm talking about the total from 1,250 -- not total from 500, at this point, to make sure we don't have confusion on the two numbers -- aerosolized. Now, of that two percent, one percent of the two percent was vapor; one percent was droplets. The way we were able to test that at Dugway was setting up sensors on the towers. We had cards placed all on the desert floor. In fact if you remember, at one point we set up a test where the media was allowed to come if they'd like to see it, and then it didn't come off. One of the reasons it didn't happen that night was the wind was blowing the opposite direction of where our cards were laid down. The whole purpose of having the cards down was to gather the data of these droplets and so on, to get that understanding. So that's how we were able to walk through that.
Through the evaporative tests that Bernie had mentioned before -- both from the wood and from the soil, we were able to determine what evaporation rate came from those and from the tests at Dugway determined what percentage went into the wood in the crates, what percentage spilled into the soil. So we were able to lay out all those percentages.
Before any of the Dugway tests, it was just guesswork on that. That's why the Dugway test was so important for bounding that source term.
Q: The cloud produced from the detonation contains what? What went up, what came down? Were you able to determine that, as to the potential for harm to those on the ground?
Mr. Walpole: There are two components to making up the cloud. One is what is aerosolized in the immediate event. That went up as vapor and droplets, droplets that later fell. The other contributor to that cloud is the evaporation over the two to three day period of time, primarily two to three day period of time, from what was spilled into the wood and the soil, so that was an ongoing evaporation -- two contributors. That's why you've got a moving cloud.
I think you wanted to add to that...?
Dr. Rostker: Yeah. We had talked about the fact that it had rained on the 11th. Those were scattered thunderstorms and we took no credit for that in terms of hydrolization. So since we couldn't pin that down where the storms were relative to the cloud, we simply ignored the fact in this case in the direction of being conservative, that there would have been local hydrolization or neutralization of the Sarin from mixture with water.
Q: What came down as aerosol is not going to stay in the sand, it's going to bake out with the sun, it's going to break down?
Mr. Walpole: Yeah, it's going to degrade. One of the issues where we don't have a lot of data, at least for the model, is what the degradation rate would be over those two to three days in the model. So our modelers didn't run degradation, another conservative point that we have in the model. But once the model is through running, to the extent that it runs, then that's disbursed at that point, it's down on the ground, either fallen as droplets or as vapor. Well, at that point it's going to be condensed into the soil, so it's going to be droplets.
Q: I don't understand what's in the report. Do you wind up not using any of the Livermoore models?
Dr. Rostker: Yes, we did not use the Livermoore models. This was a matter of just the time constraints that we had. We had made a commitment to the President's Advisory Committee to report the week of the 21st of July and we had to use, we used the resources that were available to us within the DoD, and you'll note that Livermoore is just one of several models, the models that we chose were endorsed by the IDA group...
Q: Isn't the Livermoore model, especially Matthew, much more highly precise based on...
Dr. Rostker: It's not my understanding.
We have a representative from IDA here.
Mr. Grotti: Good afternoon. I'm Jeff Grotti from the Institute for Defense Analyses. I'm listed on the cover letter as the project leader.
We looked at the models, the Department of Defense models and the energy lab models; or I should say our meteorologists looked at them very carefully and the results that they were producing, and we had them produce other results. We actually felt that the results that they were producing was not consistent with what the other models were looking at. In fact the other models the Department of Defense models, do a better job of modeling the atmospheric physics than the Matthew model does. For that reason in the report we did not recommend that those models be used.
Q: What do you think was the most valuable contribution that the IDA study specifically made to all of this? And can you also tell me how much the contract was for, dollar-wise?
Mr. Grotti: Yes. The contract that IDA got was for $150,000 to cover the panel activities. There was additional funds, and I apologize for not having the number that was provided not from IDA but through the Department of Defense, to the various groups of modelers to do additional work for us so that we could understand the models better and understand some of the sensitivities, the unknowns, that we were very concerned about.
I think the contribution of the IDA study is the methodology that we have forwarded to the Department of Defense. As Bob summed it up, to use multiple models, to link the models, to try and get a handle on the variations and so forth. To try and ensure that all those are taken into account when you try and generate the coverage of this agent -- especially at low levels, which is difficult to do.
Q: I don't want to press, I'll just ask you real quick. The reason that you rejected the Livermoore model is because you felt it was not appropriate to the circumstance, or do you think their science is flawed in establishing the kind of modeling they do?
Mr. Grotti: I don't think their science is flawed. Of course they have this airborne release advisory capability, it's been used widely. We just feel that in this case, because of the lack of data, local data. There's a lot of global data that was available but there wasn't a lot of data about the meteorology in the area. The other models were able to make up for that because they had the better physics and they took into account the fact that there were bodies of water and the terrain, and there were swamps here and so forth, and they were able to take all that into account. The Matthew model was a little less able to take that into account.
Q: I thought it was more scientifically precise though. No? Okay.
Mr. Grotti: Precise is not accurate, is a truism of science. You get good results, but they may not be the right results.
In this case, and I don't want to extend this observation to anything else they might be doing. I've got to say they really did a nice job of helping us understand their models and they contributed a great deal to the panel, but the panel was just not comfortable with those results for this particular instance.
Dr. Rostker: As we were finishing doing some of our runs trying to put all this together, we also enlisted the consultation of Dr. Bruce Hicks from NOAA's Atmospheric Research Laboratory. Bruce had not been exposed to any of this, and he came in as a last minute guru. He expressed to us the fact that he felt we were breaking new ground, particularly in bringing the models together, multiple models. We talked about an ensemble of models, and so we've had, again, further outside consultation. But ultimately, these are all simulation models and they represent our best assessment as to where this cloud may have gone, what its density was, what dosages people under the clouds would have been. These are still approximations.
For example, we have in our data point estimate for units basically one observation per day. We know these units were moving, and we know that not every individual on the troop list for the unit was with the precise center of the unit. So all this still represents approximation. But I would emphasize that we are, as you will see next week, substantially talking about low level exposures all the way out to the population limits of 72 hours of continuous exposure in the general population. The area of where there are symptoms you will see is a very small area.
One of the things we'll do next week also is to plot for you the area that our alarms would have gone off. One of the issues has always been why didn't the alarms go off. We'll plot the area where there was sufficient concentration to set the alarms off.
We also have some very preliminary data in the sense of we had questionnaires to the 20,000 people around Khamisiyah and we asked those people about whether they thought they might have been ill because of exposure to chemicals. There were some 370 people out of the 20,000 questionnaires that responded to that. There were about 7,000 responses and there were about 370, if memory serves, that responded to that. We have attempted to call every one of them, and we have spoken to well over 50 percent, and we were able to then eliminate some who said well, I really wasn't sick, there was a coding error and the like. Then we have some symptoms. The symptoms are not unique to potential chemical exposure. The symptoms include rash, diarrhea, headaches, but we're able to take the people and look at where they were relative to other people in their unit. You'll see, again we're talking about onesies and twosies relative to the population, so we will be looking at those data. Subsequent to this we will be, not for next week, we will be sorting the plume with people who are in the CCEP or in the VA Registry. The Navy health research laboratory in San Diego has already started analysis of plume rates and participation rates in the registry based upon distance from Khamisiyah and the like, and we can now give them much better indications of who were there and what we believe to be the fallout pattern.
Q: Had you had any estimate prior to this with the number of rockets in the pit?
Mr. Walpole: We have had multiple efforts to try to figure out the number of rockets in the pit. In fact I think in our Khamisiyah paper, if we didn't state it in there, at the time we were writing the Khamisiyah paper we had an estimate of about 1,400, which is our current upper bound. The lower bound for the estimate is 1,100 and that's an easy lower bound. That's the amount the Iraqis declared to be in the pit. So we know that we didn't have less than what they said was there, we just don't happen to believe that that's as few as were there.
We did a lot of ways to look at that. In interviewing the soldiers as they described which stack -- remember there were 13 stacks there. They would describe how high the stack was relative to their height. Then in doing measuring of photography, trying to calculate all that, we have a much better handle now and feel that 1250 is a much better estimate.
Q: How many rockets in the pit may not have been destroyed? Based on your Dugway test.
Mr. Walpole: We have even better information than that. UNSCOM had 744 -- we rounded 750 to keep the numbers round here -- that were not affected. When they went to the pit they counted that many that they had to pull aside and destroy. So the 500 number that's in the paper here, that's the one that make up the other 40 percent of what went into the wood, what went into the soil, what was burned and what was aerosolized.
Q: So everything that was in the pit's been destroyed.
Mr. Walpole: Has been accounted for.
Q: Is it buried in Khamisiyah?
Mr. Walpole: No. It's been since destroyed. UNSCOM destroyed those other 744. Is that what you mean?
Dr. Rostker: UNSCOM blew up a number of rockets, damaged rockets, leaking rockets on the scene, and we have videotapes of UNSCOM doing all of that. The rockets that were intact were carted off by UNSCOM and were incinerated.
Q: Based on what you know now, a week away from releasing your estimate, are you convinced that at least some number of U.S. troops was exposed to low levels of chemicals as a result of this?
Dr. Rostker: Yes.
Q: When you release this estimate, is it going to be a big? What kind of a fudge factor is it going to have? How confident are you going to be in the number that you're going to be able to release next week?
Dr. Rostker: We're confident in the work we've done in Dugway. We're confident in our modeling. It's conservative. We think it's good. I don't have a number for you now and I don't even want to speculate. We will be here next Thursday and...
Q: Are you going to be telling us when you give us this number that it's a guesstimate?
Dr. Rostker: No. I'm going to tell you it is the best number that we can provide you after studying this intensely. A lot of people studying this intensely for a year.
Q: Just to be clear, when you release this estimate, it's not going to answer the question of whether that low level exposure might account for the illnesses and ailments suffered by Gulf War veterans?
Dr. Rostker: This is part of a broad puzzle. We are providing information to that to try to solve that puzzle, but this is still a specific event in time. It's still a specific event in distance, in geography. There are people who have illness that were not in the Gulf at that time and were not in any area that we believe could have had the fallout. So this is part of the puzzle. It will help--and this is important--this is not just an exercise to notify people. This is an exercise to get as good a number as we can, because we think it's an important contribution to subsequent research to answer the questions you have, Jamie. So we've strived not just to draw arbitrary lines and send letters out, we've strived to get as good a number as we can have so that as we do subsequent research we'll be able to understand the potential for chemical, low level chemicals that might have on the health of the people who were in the Gulf.
Q: You're saying once you get the shape and the drift of the cloud it will allow you to better pinpoint your investigation of complaints, and they are...
Dr. Rostker: We will be able to then correlate it with health data, with questionnaire data, with the various studies that are going on. You know, all of this has always been classic: You have a treatment, you look for the effect. Well, we've never understood what the treatment was -- who was exposed or who was not exposed. This gives us much more precision in understanding that for subsequent research.
Press: Thank you.