Science Inventory

You are here:

EPA Home » Science Inventory » UNCERTAINTY AND SENSITIVITY ANALYSES FOR VERY HIGH ORDER MODELS

UNCERTAINTY AND SENSITIVITY ANALYSES FOR VERY HIGH ORDER MODELS

Contact

http://cfpub.epa.gov/si/si_public_comments.cfm

Impact/Purpose:

The primary goals are to: (1) Construct a 400-node PC-based supercomputing cluster supporting Windows and Linux computer operating systems (i.e. SuperMUSE: Supercomputer for Model Uncertainty and Sensitivity Evaluation); (2) Develop platform-independent system software for the management of SuperMUSE and parallelization of EPA models and modeling systems for implementation on SuperMUSE (and other PC-based clusters); (3) Conduct uncertainty and sensitivity analyses of the 3MRA modeling system; (4) Develop advanced algorithmic software for advanced statistical sampling methods, and screening, localized, and global sensitivity analyses; and (5) Provide customer-oriented model applications for probabilistic risk assessment supporting quality assurance in multimedia decision-making.

Description:

While there may in many cases be high potential for exposure of humans and ecosystems to chemicals released from a source, the degree to which this potential is realized is often uncertain. Conceptually, uncertainties are divided among parameters, model, and modeler during simulation, where inaccuracy in model predictions results principally from lack of knowledge and data across this realm. In comparison, sensitivity analysis can lead to a better understanding of how models respond to variation in their inputs, which in turn can be used to better focus laboratory and field-based data collection efforts on processes and parameters contributing most to uncertainty in outcomes of interest. There is a general need to both describe uncertainty for the current state of science and data, and, further, to ascertain a prioritized agenda for the reduction of this uncertainty. This must be facilitated across a broad suite of application purposes for models of widely varying complexity, each with varying dimensionality needed to describe variability in time, space, and population membership.

Uncertainty analysis allows for the critical task of making informed decisions in the present, based on accumulated knowledge and data. This is complemented by the many powers of sensitivity analysis to inform directions for both data collection and model improvement for specific purposes. Quantitative analysis of integrated, very high order multimedia models will ultimately require a comparative model evaluation approach using several algorithmic techniques, coupled with sufficient computational power.

Residing within the Framework for Risk Analysis in Multimedia Environmental Systems (FRAMES), the Multimedia, Multipathway, and Multireceptor Risk Assessment (3MRA) modeling system is an example, very high order modeling system being developed by EPA for use in assessment of hazardous waste management facilities. 3MRA currently encompasses 966 variables, over 185 of which are stochastic. A characteristic of uncertainty analysis (UA) and sensitivity analysis (SA) for very high order models (VHOMs) like 3MRA is their need for significant computational capacity to perform relatively redundant simulations. While UA/SA is emerging as a critical area for environmental model evaluation, resources for Windows-based models have been limited by an associated lack of supercomputing capacity. Equally, higher-order UA/SA algorithms warrant investigation to determine their efficacy in establishing requisite confidence in the use of VHOMs for regulatory decision-making.

This task delivers a PC-based, Windows-based supercomputing capability to facilitate UA/SA research. It also provides short-term and long-term plans for development and integration of a robust, functional UA/SA tool set within FRAMES (and other frameworks) for evaluation of VHOMs and simpler models, and associated applications.

Record Details:

Record Type:PROJECT
Start Date:10/01/2003
Projected Completion Date:09/01/2008
OMB Category:Other
Record ID: 56122