policy guidance on the implementation of QMU from NNSA, will be essential for improving consistency and transparency in the implementation of QMU and for facilitating peer review. Accordingly, documentation must be given high priority.
As noted above, the use of different approaches to QMU can be a strength as long as methods are documented to make them more transparent and to assist researchers in communicating effectively with one another, with management, and with outside audiences.
Recommendation 3-3. It is urgent that NNSA and the national security labs complete the development and issuance of QMU guidance documents in time for the current assessment cycle. This process should be used to drive consensus among lab scientists. The documents should be updated as the methodology matures.
Finding 3-4. The QMU framework has yet to be clearly defined by the national security laboratories collectively or individually. This framework must identify a more comprehensive set of performance gates and describe how QMU is used to analyze each. A possible outcome of this process is that QMU is not appropriate for a particular performance gate.
QMU is often conflated with the whole set of tasks and tools that must be carried forward for stockpile stewardship and design of the RRW. These tasks and tools exist independently of the way that QMU is defined. The tools used in the QMU process are, for the most part, already in widespread use; this is not the issue. Rather, the issue is that the overarching QMU process needs to take into account various divergent views on the essence of the process.
An incomplete QMU methodology could also result in a situation in which the blind application of QMU increases the likelihood of missing an alternative failure mechanism or of hiding it altogether. If this happened, efforts to increase a margin and improve the apparent confidence factor of a nuclear explosive package determined from the application of QMU could activate an alternative failure mechanism. For example, design changes that enhance yield margin could introduce one-point-safety concerns.
Recommendation 3-4. The national security labs should carry out interlaboratory comparisons of different methods for finding and characterizing the most important uncertainties and for propagating these uncertainties through computer simulations of weapons performance.