Towards Long-Term Scientific Model Sustainment at Sandia National Laboratories

Software quality assurance in the 1996 performance assessment for the Waste Isolation Pilot Plant

The Edwards aquifer is a karstic aquifer which discharges naturally to springs that support a number of endangered aquatic species. Past and projected rates of development and urbanization are high as are the political stakes in managing regional growth and water resources in a sustainable manner. Key issues involve how can development occur or what type of development is acceptable to maintain critical spring flows, water quality, maximum pumping from the aquifer, and environmental quality with intergenerational equity. We are developing a decision support system (DSS) that links scientific and management models that incorporate community stakeholder input, optimization algorithms, and decision analysis tools. The Texas Water Development Board's groundwater availability model is the basis for the DSS. We scale the distributed parameter model (MODFLOW) into a systems dynamics model (POWERSIM © ) with links to socio-economic data and stakeholder preferences. Key decision variables include: well distribution and pumping rates, impervious cover, water and sewage line distributions, and stream setbacks. Key outputs include spring flow and water table declines in drought years. Overcoming uncertainties in the relationships between land use, recharge, and aquifer performance is a significant challenge. Understanding the interconnection between systems and hypothesis testing to bracket expected changes in recharge provide valuable information for groundwater management, which we are testing through stakeholder input, including narrative elicitation. The DSS allows comparison of hypotheses about spatial recharge distributions and communication with decision makers in a timely, policy-relevant format that is compatible with Texas policy for quantifying consensus and available yields. This research is being conducted at The University of Texas at Austin and Sandia National Laboratories.

Uncertainty quantification (UQ) is the study of the effects of uncertainty on the values of analytical results and the predictions of scientific models. Sources of uncertainty include imprecise knowledge of the exact values of parameters, lack of confidence in the physical models, use of imperfectly calibrated models, and irreducible uncertainties due to physical characteristics. The Air Force Research Laboratory has undertaken the challenge of understanding, developing and analyzing the techniques of UQ as they apply to Laser Beam Control. This paper proposes a simple methodology and simple results with our first attempt of applying UQ as a new analysis tool. The software toolkit which was chosen was an analytical group of algorithms from a Sandia National Laboratory (SNL) package called DAKOTA (Design Analysis Kit for Optimization and Terascale Applications). The specific application of interest to the Air Force Research Laboratory (AFRL) is the analytical prediction of the performance of a Laser Beam Control systems under various scenarios, conditions, and missions. The application of rigorous UQ techniques to the models used to predict beam control performance could greatly improve our confidence in these predictions and also improve the acceptance of advanced Laser Beam Control systems within the science and engineering communities1,2. The proposed work would follow a multi-step approach, analyzing the more easily quantified sources of uncertainty, and then including increasingly complicated physical phenomena as the work progresses. Will present the initial results, and the first steps in the incorporation of UQ into our Laser Beam Control Modeling and Simulation environments.

Computational environment and software configuration management of the 1996 performance assessment for the Waste Isolation Pilot Plant