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

Title: A geostatistics-informed hierarchical sensitivity analysis method for complex groundwater flow and transport modeling: GEOSTATISTICAL SENSITIVITY ANALYSIS

Authors:
ORCiD logo [1]; ORCiD logo [1]; ORCiD logo [2]; ORCiD logo [1];  [1]
  1. Pacific Northwest National Laboratory, Richland Washington USA
  2. Department of Scientific Computing, Florida State University, Tallahassee Florida USA
Publication Date:
Sponsoring Org.:
USDOE Office of Science (SC), Biological and Environmental Research (BER) (SC-23)
OSTI Identifier:
1402249
Grant/Contract Number:
SC0008272; 1552329; AC02-05CH11231
Resource Type:
Journal Article: Publisher's Accepted Manuscript
Journal Name:
Water Resources Research
Additional Journal Information:
Journal Volume: 53; Journal Issue: 5; Related Information: CHORUS Timestamp: 2017-10-23 17:24:33; Journal ID: ISSN 0043-1397
Publisher:
Wiley Blackwell (John Wiley & Sons)
Country of Publication:
United States
Language:
English

Citation Formats

Dai, Heng, Chen, Xingyuan, Ye, Ming, Song, Xuehang, and Zachara, John M. A geostatistics-informed hierarchical sensitivity analysis method for complex groundwater flow and transport modeling: GEOSTATISTICAL SENSITIVITY ANALYSIS. United States: N. p., 2017. Web. doi:10.1002/2016WR019756.
Dai, Heng, Chen, Xingyuan, Ye, Ming, Song, Xuehang, & Zachara, John M. A geostatistics-informed hierarchical sensitivity analysis method for complex groundwater flow and transport modeling: GEOSTATISTICAL SENSITIVITY ANALYSIS. United States. doi:10.1002/2016WR019756.
Dai, Heng, Chen, Xingyuan, Ye, Ming, Song, Xuehang, and Zachara, John M. Wed . "A geostatistics-informed hierarchical sensitivity analysis method for complex groundwater flow and transport modeling: GEOSTATISTICAL SENSITIVITY ANALYSIS". United States. doi:10.1002/2016WR019756.
@article{osti_1402249,
title = {A geostatistics-informed hierarchical sensitivity analysis method for complex groundwater flow and transport modeling: GEOSTATISTICAL SENSITIVITY ANALYSIS},
author = {Dai, Heng and Chen, Xingyuan and Ye, Ming and Song, Xuehang and Zachara, John M.},
abstractNote = {},
doi = {10.1002/2016WR019756},
journal = {Water Resources Research},
number = 5,
volume = 53,
place = {United States},
year = {Wed May 31 00:00:00 EDT 2017},
month = {Wed May 31 00:00:00 EDT 2017}
}

Journal Article:
Free Publicly Available Full Text
This content will become publicly available on May 31, 2018
Publisher's Accepted Manuscript

Save / Share:
  • Sensitivity analysis is an important tool for quantifying uncertainty in the outputs of mathematical models, especially for complex systems with a high dimension of spatially correlated parameters. Variance-based global sensitivity analysis has gained popularity because it can quantify the relative contribution of uncertainty from different sources. However, its computational cost increases dramatically with the complexity of the considered model and the dimension of model parameters. In this study we developed a hierarchical sensitivity analysis method that (1) constructs an uncertainty hierarchy by analyzing the input uncertainty sources, and (2) accounts for the spatial correlation among parameters at each level ofmore » the hierarchy using geostatistical tools. The contribution of uncertainty source at each hierarchy level is measured by sensitivity indices calculated using the variance decomposition method. Using this methodology, we identified the most important uncertainty source for a dynamic groundwater flow and solute transport in model at the Department of Energy (DOE) Hanford site. The results indicate that boundary conditions and permeability field contribute the most uncertainty to the simulated head field and tracer plume, respectively. The relative contribution from each source varied spatially and temporally as driven by the dynamic interaction between groundwater and river water at the site. By using a geostatistical approach to reduce the number of realizations needed for the sensitivity analysis, the computational cost of implementing the developed method was reduced to a practically manageable level. The developed sensitivity analysis method is generally applicable to a wide range of hydrologic and environmental problems that deal with high-dimensional spatially-distributed parameters.« less
  • The overall theme of the DOE/AECL '87 conference was the application of statistical methods for sensitivity and uncertainty analysis to nuclear waste disposal flow and transport modeling. The conference was organized into six technical sessions dealing with different topics in that theme: the role of statistical methods in the nuclear waste repository performance assessment plans of several waste disposal programs, sensitivity and uncertainty methods for large computer codes, stochastic hydrology, statistical methods for modeling flow and transport in fractured media, geostatistical methods, and the use of subjective information to supplement hard data in performance assessment studies. Thirty-one presentations were mademore » in the six sessions. Several presentations were invited from well-known authorities in the various technical fields. These invited presentations were supplemented by the contributed presentations of researchers from seven different countries. At the combustion of the technical sessions, an open discussion session was held to address areas of concern to all conference attendees.« less
  • It is challenging to predict the degree to which shallow groundwater might be affected by leaks from a CO 2 sequestration reservoir, particularly over long time scales and large spatial scales. In this study observations at a CO 2 enriched shallow aquifer natural analog were used to develop a predictive model which is then used to simulate leakage scenarios. This natural analog provides the opportunity to make direct field observations of groundwater chemistry in the presence of elevated CO 2, to collect aquifer samples and expose them to CO 2 under controlled conditions in the laboratory, and to test themore » ability of multiphase reactive transport models to reproduce measured geochemical trends at the field-scale. The field observations suggest that brackish water entrained with the upwelling CO 2 are a more significant source of trace metals than in situ mobilization of metals due to exposure to CO 2. The study focuses on a single trace metal of concern at this site: U. Experimental results indicate that cation exchange/adsorption and dissolution/precipitation of calcite containing trace amounts of U are important reactions controlling U in groundwater at this site, and that the amount of U associated with calcite is fairly well constrained. Simulations incorporating these results into a 3-D multi-phase reactive transport model are able to reproduce the measured ranges and trends between pH, pCO 2, Ca, total C, U and Cl -at the field site. Although the true fluxes at the natural analog site are unknown, the cumulative CO 2 flux inferred from these simulations are approximately equivalent to 37.8E-3 MT, approximately corresponding to a .001% leak rate for injection at a large (750 MW) power plant. The leakage scenario simulations suggest that if the leak only persists for a short time the volume of aquifer contaminated by CO 2-induced mobilization of U will be relatively small, yet persistent over 100 a.« less
  • Vertical and horizontal transects were sampled from core and outcrop of the San Andres Formation at Lawyer Canyon, Guadalupe Mountains, New Mexico, to assess permeability variation in a geologic framework of upward-shallowing carbonate cycles and to show the potential effect these variations have on viscous-dominated flow behavior in analogous reservoirs. These cycles occur in a ramp-crest facies, tract, are 3-13 m (10-45 ft) thick, and contain both vertical and lateral variation of lithofacies. Thicker cycles consist of a basal dolomudstone, which is overlain by burrowed dolomudstone, and capped by bar-flank ooid-peloid dolograinstone and bar-crest ooid dolograinstones. In vertical transects, permeabilitymore » is extremely variable about the mean, yet upward-increasing trends coinciding with the succession of lithofacies typify a given cycle. Semi-variance analysis shows permeability to be uncorrelated vertically at distances greater than 5.5 m (18 ft), which is the average cycle thickness, suggesting that the cycles may equate to fluid-flow unit in a reservoir. Semi-variance analysis of measurements collected along a horizontal transect within bar-crest dolograinstones of a single cycle show permeability is uncorrelated at distances greater than 3.6 m (12 ft). This correlation distance appears to be controlled by alternating porous and tightly cemented zones that formed during dolomitization. Vertical and lateral variogram models were fit to the spatial parameters to generate a variety of conditionally simulated permeability fields. Fluid-flow simulations show viscous-dominated flow behavior is compartmentalized by both the individual cycles and groups of cycles. The basal dolomudstones are potential baffles to flow crossover between cycles, but poorly developed cycles (i.e., those that are mud rich and lack well-developed bar-flank and bar-crest facies) result in the greatest compartmentalization of fluid flow within a succession of cycles.« less
  • A multivariate Analysis of Variance (ANOVA) is used to measure the relative sensitivity of groundwater flow to two factors that indicate different dimensions of aquifer heterogeneity. An aquifer is modeled as the union of disjoint volumes, or blocks, composed of different materials with different hydraulic conductivities. The factors are correlation between the hydraulic conductivities of the different materials and the contrast between mean conductivities in the different materials. The precise values of aquifer properties are usually uncertain because they are only sparsely sampled, yet are highly heterogeneous. Hence, the spatial distribution of blocks and the distribution of materials in blocksmore » are uncertain and are modeled as stochastic processes. The ANOVA is performed on a large sample of Monte Carlo simulations of a simple model flow system composed of two materials distributed within three disjoint blocks. Our key finding is that simulated flow is much more sensitive to the contrast between mean conductivities of the blocks than it is to the intensity of correlation, although both factors are statistically significant. The methodology of the experiment - ANOVA performed on Monte Carlo simulations of a multi-material flow system - constitutes the basis of additional studies of more complicated interactions between factors that define flow and transport in aquifers with uncertain properties.« less