are incomplete, or have not been analyzed. The effort to compile and evaluate these data is considerable, but there is enough new information on field performance, material behavior, and monitoring and modeling capabilities to make an assessment of performance worthwhile about every 5 to 10 years. More frequent assessments may be required based on previous monitoring data and performance assessment models.

Recommendation 3: Federal agencies responsible for engineered barrier systems should commission and fund assessments of performance on a regular basis. Given the rate at which performance data and knowledge of waste behavior, contaminant transport, and monitoring accumulate, the interval at which these assessments should take place is probably on the order of once every 5 to 10 years. The results of the assessment should be placed in the public domain in a form that is readily accessible.

Much data used to predict performance come from laboratory experiments, models, and field-constructed prototype barrier systems (e.g., test pads). Although useful for understanding material properties and behavior, these data are no substitute for performance data collected in the field from operating containment systems. An overall comprehensive assessment of performance requires long-term monitoring and analysis of data from different types of waste containment systems constructed from a variety of components and located in different climate regimes.

Recommendation 4: EPA, USNRC, NSF, and DOE should establish a set of observatories at operational containment facilities to assess the long-term performance of waste containment systems at field scale. The program would involve building one or more field facilities, monitoring the site, and analyzing and archiving the data. New sites could be created or adjustments could be made to existing observatories when promising new and innovative concepts and materials become available.

Analytical and numerical models are relied on to predict contaminant transport, containment effectiveness, degradation of materials, and changes in behavior over time, even though some models have shortcomings (e.g., they do not account for advection-dispersion processes; they are used in applications for which they were not designed).

Recommendation 5: Regulatory agencies (e.g., EPA, DOE, USNRC) and research sponsors (e.g., NSF) should support the validation, calibration, and improvement of models to predict the behavior of containment system components and the composite system over long periods of time. These models should be validated and calibrated using the results of field observations and measurements.

The optimum time for monitoring varies with the facility, type of waste, climate, and the observed performance. Yet funding is often not available to continue monitoring until the site no longer poses risk to human health and the environment, and no national policy exists to assure that such funding will be available.

Recommendation 6: EPA should develop financial assurance mechanisms to ensure that funding is available for monitoring and care for as long as the waste poses a threat to human health and the environment.

Performance criteria are needed that account for both barrier performance and impacts to public health and safety that extend beyond the barrier system.

Recommendation 7: EPA and USNRC should develop guidance for the practical implementation of performance-based criteria for assessment of containment system performance as an alternative to prescriptive designs.