skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Velocity prediction errors related to flow model calibration uncertainty

Conference ·
OSTI ID:6075987
 [1]; ;  [2]
  1. Westinghouse Savannah River Co., Aiken, SC (USA)
  2. Geraghty and Miller, Inc., Reston, VA (USA)

At the Savannah River Site (SRS), a United States Department of Energy facility in South Carolina, a three-dimensional, steady-state numerical model has been developed for a four aquifer, three aquitard groundwater flow system. This model has been used for numerous predictive simulation applications at SRS, and since the initial calibration, the model has been refined several times. Originally, calibration of the model was accomplished using a nonlinear least-squares inverse technique for a set of 50 water-level calibration targets non-uniformly distributed in the four aquifers. The estimated hydraulic properties from this calibration generally showed reasonable agreement with values estimated from field tests. Subsequent model refinements and application of this model to field problems have shown that uncertainties in the model parameterization become much more apparent in the prediction of the velocity field than in the simulation of the distribution of hydraulic heads. The combined use of three types of information (hydraulic head distributions, geologic framework models, and velocity field monitoring) provide valuable calibration data for flow modeling investigations; however, calibration of a flow model typically relies upon measured water levels. For a given set of water-level calibration targets, the uncertainties associated with imperfect knowledge of physical system parameters or groundwater velocities may not be discernable in the calibrated hydraulic head distribution. In this paper, modeling results from studies at SRS illustrate examples of model inadequacy resulting from calibrating only on observed water levels, and the effects of these inadequacies on velocity field prediction are discussed. 14 refs., 6 figs.

Research Organization:
Westinghouse Savannah River Co., Aiken, SC (USA)
Sponsoring Organization:
DOE/DP
DOE Contract Number:
AC09-89SR18035
OSTI ID:
6075987
Report Number(s):
WSRC-MS-90-124; CONF-9009254-2; ON: DE91005651
Resource Relation:
Conference: International conference on calibration and reliability in groundwater modeling, Hague (Netherlands), 3-6 Sep 1990
Country of Publication:
United States
Language:
English