The Panel, members of which are listed in the Appendix, met twice once on 18- 20
November, 1996 and again on 13 February, 1997 and communicated extensively between
meetings and after the February meeting. At the first meeting, the Panel reviewed what was
then known about the demolition in the pit, the analyses performed for the CIA by SAIC,
and modeling capabilities extant in the Department of Defense and the Department of
Energy. The second meeting was held to review additional analyses that had been requested
at the first meeting. This report summarizes the findings of the Panel.
The CIA and their contractor reported to the Panel unpublished analyses of their analysis of the pit demolition. Their briefings detailed a number of assumed parameters, such as the number of rockets that could have been affected by the demolition, the number that may have been flyouts, agent purity and other characteristics, and the mechanisms by which agent may have been released.
The SAIC analysis utilized a number of transport and diffusion models: NUSSE4 and D2PC, which are maintained by the US Army Chemical and Biological Defense Command; VLSTRACK, which is maintained by the Navy Surface Warfare Center; and the OMEGA mesoscale atmospheric prediction model, which is being developed by the Defense Special Weapons Agency. Based on the briefings received, the Panel assessed that SAIC primarily used the NUSSE4 model, which assumed a constant expected value of meteorological conditions, to determine the farthest downwind direction that agent might travel under a variety of release conditions; the OMEGA model to detenrune the directions that local winds may have carried agent; and the D2PC model, which is used by DOD to model potential releases from the US chemical stockpile and demilitarized sites, to represent the spill source. Although SAIC briefed that VLSTRACK and OMEGA were linked, and that analytical linkages were then made to NUSSE4 and D2PC, it is the view of the Panel that these meteorological, source, and transport and diffusion models were not adequately linked.
The outputs of these models, as interpreted by the Panel, were combined to yield an "area of concern"-a sector, the subtended angle of which was determined by the extent to which winds could have moved agent over a 48 hour period, and the radius of which was determined by the distance that the dispersion model predicted the agent would persist above a specific dosage level, such as the level specified as safe for the general population in Army material safety publications.
The CIA suspended the pit analysis in September, 1996, when it became clear that the uncertainties were too great to continue. Not only were the source factors that were reflecting the phenomenology of this release not well understood, but also the observations supporting the meteorological reconstruction were very sparse around Khamisiyah.