Parameter estimation and predictive uncertainty in stochastic inverse modeling of groundwater flow: Comparing null‐space Monte Carlo and multiple starting point methods

[1] Given a highly parameterized groundwater model in which the conceptual model of the heterogeneity is stochastic, a set of inverse calibrations from multiple starting points (MSPs) provide an ensemble of calibrated parameters and follow-on transport predictions. However, the multiple calibrations are computationally expensive. A recently developed null-space Monte Carlo (NSMC) method combines the calibration solution-space parameters with the ensemble of null-space parameters, creating sets of calibration-constrained parameters for input to follow-on transport predictions. The consistency between probabilistic ensembles of parameter estimates and predictions created using the MSP calibration and the NSMC approaches is examined using a highly parameterized (>1300 parameters) model of the Culebra dolomite previously developed for the Waste Isolation Pilot Plant project in New Mexico as a test case. A total of 100 estimated fields are retained from the MSP approach, and the ensemble of results defining the model fit to the data and prediction of an advective travel time are compared with the same results obtained using NSMC. We demonstrate that the NSMC fields based on a single calibrated model can be significantly constrained by the calibrated solution space, and the resulting distribution of advective travel times is biased toward the travel time from the single calibrated field. To overcome this, newly proposed strategies to employ a multiple calibration-constrained NSMC (M-NSMC) approach are evaluated. Comparison of the M-NSMC and MSP methods demonstrates that M-NSMC can provide a computationally efficient and practical solution for predictive uncertainty analysis in highly nonlinear and complex subsurface flow and transport models.

[1]  G. Dagan,et al.  Stochastic identification of transmissivity and effective recharge in steady groundwater flow: 2. Case study , 1987 .

[2]  G. Marsily,et al.  Regards sur 40 ans de problèmes inverses en hydrogéologie , 1999 .

[3]  Andres Alcolea,et al.  Regularized pilot points method for reproducing the effect of small scale variability : Application to simulations of contaminant transport , 2008 .

[4]  P. Kitanidis.  On the geostatistical approach to the inverse problem , 1996 .

[5]  Y. Rubin,et al.  Investigation of flow and transport in certain cases of nonstationary conductivity fields , 1994 .

[6]  John Doherty,et al.  Efficient regularization and uncertainty analysis using a global optimization methodology , 2010 .

[7]  John Doherty,et al.  Approaches to Highly Parameterized Inversion: A Guide to Using PEST for Model-Parameter and Predictive-Uncertainty Analysis , 2014 .

[8]  S. P. Neuman,et al.  Maximum likelihood Bayesian averaging of uncertain model predictions , 2003 .

[9]  D. Kolpin,et al.  A national reconnaissance of pharmaceuticals and other organic wastewater contaminants in the United States--I) groundwater. , 2008, The Science of the total environment.

[10]  Brian E. Skahill,et al.  More efficient PEST compatible model independent model calibration , 2009, Environ. Model. Softw..

[11]  Q. Kang,et al.  Optimization and uncertainty assessment of strongly nonlinear groundwater models with high parameter dimensionality , 2010 .

[12]  Hongkyu Yoon,et al.  Highly parameterized inverse estimation of hydraulic conductivity and porosity in a three‐dimensional, heterogeneous transport experiment , 2012 .

[13]  L. Hu.  Gradual Deformation and Iterative Calibration of Gaussian-Related Stochastic Models , 2000 .

[14]  John Doherty,et al.  Approaches to Highly Parameterized Inversion: Pilot-Point Theory, Guidelines, and Research Directions , 2014 .

[15]  Mary C. Hill,et al.  Comment on ``Two statistics for evaluating parameter identifiability and error reduction'' by John Doherty and Randall J. Hunt , 2010 .

[16]  F. Pappenberger,et al.  Ignorance is bliss: Or seven reasons not to use uncertainty analysis , 2006 .

[17]  Keith Beven,et al.  The future of distributed models: model calibration and uncertainty prediction. , 1992 .

[18]  John Doherty,et al.  Ground Water Model Calibration Using Pilot Points and Regularization , 2003, Ground water.

[19]  Yoram Rubin,et al.  Field investigation of the waste isolation pilot plant (WIPP) site (New Mexico) using a nonstationary stochastic model with a trending hydraulic conductivity field , 1999 .

[20]  M. G. Marietta,et al.  Pilot Point Methodology for Automated Calibration of an Ensemble of Conditionally Simulated Transmissivity Fields: 2. Application , 1995 .

[21]  M. Boucher,et al.  Interpretation of Interference Tests in a Well Field Using Geostatistical Techniques to Fit the Permeability Distribution in a Reservoir Model , 1984 .

[22]  S. P. Neuman,et al.  Estimation of Aquifer Parameters Under Transient and Steady State Conditions: 1. Maximum Likelihood Method Incorporating Prior Information , 1986 .

[23]  M. Marietta,et al.  Pilot Point Methodology for Automated Calibration of an Ensemble of conditionally Simulated Transmissivity Fields: 1. Theory and Computational Experiments , 1995 .

[24]  George Kuczera,et al.  Understanding predictive uncertainty in hydrologic modeling: The challenge of identifying input and structural errors , 2010 .

[25]  John Doherty,et al.  Two statistics for evaluating parameter identifiability and error reduction , 2009 .

[26]  T. Ulrych,et al.  A full‐Bayesian approach to the groundwater inverse problem for steady state flow , 2000 .

[27]  Ghislain de Marsily,et al.  Three‐dimensional interference test interpretation in a fractured aquifer using the Pilot Point Inverse Method , 2001 .

[28]  D. A. Zimmerman,et al.  A comparison of seven geostatistically based inverse approaches to estimate transmissivities for modeling advective transport by groundwater flow , 1998 .

[29]  Keith Beven,et al.  A manifesto for the equifinality thesis , 2006 .

[30]  S. P. Neuman,et al.  Inverse stochastic moment analysis of steady state flow in randomly heterogeneous media , 2006 .

[31]  J. Doherty,et al.  Calibration‐constrained Monte Carlo analysis of highly parameterized models using subspace techniques , 2009 .

[32]  John Doherty,et al.  Efficient nonlinear predictive error variance for highly parameterized models , 2006 .

[33]  J. Doherty,et al.  A hybrid regularized inversion methodology for highly parameterized environmental models , 2005 .

[34]  Charles F. Harvey,et al.  Mapping Hydraulic Conductivity: Sequential Conditioning with Measurements of Solute Arrival Time, Hydraulic Head, and Local Conductivity , 1995 .

[35]  Andrés Sahuquillo,et al.  Stochastic simulation of transmissivity fields conditional to both transmissivity and piezometric head data—3. Application to the Culebra formation at the Waste Isolation Pilot Plan (WIPP), New Mexico, USA , 1998 .

[36]  S. Sorooshian,et al.  Effective and efficient global optimization for conceptual rainfall‐runoff models , 1992 .

[37]  John Doherty,et al.  Predictive uncertainty analysis of a saltwater intrusion model using null‐space Monte Carlo , 2011 .

[38]  G. Seber,et al.  Nonlinear Regression: Seber/Nonlinear Regression , 2005 .

[39]  John Doherty,et al.  Response to Comment on “Two statistics for evaluating parameter identifiability and error reduction” , 2010 .

[40]  John Doherty,et al.  Non-uniqueness of inverse transmissivity field calibration and predictive transport modeling , 2003 .

[41]  Barbara S. Minsker,et al.  Incorporating subjective and stochastic uncertainty in an interactive multi-objective groundwater calibration framework , 2010 .

[42]  Bryan A. Tolson,et al.  Dynamically dimensioned search algorithm for computationally efficient watershed model calibration , 2007 .

[43]  Willem A. Schreuder,et al.  Using a Cloud to Replenish Parched Groundwater Modeling Efforts , 2010, Ground water.

[44]  J. Doherty,et al.  Role of the calibration process in reducing model predictive error , 2005 .

[45]  R. Hunt,et al.  Assessment of sewer source contamination of drinking water wells using tracers and human enteric viruses. , 2010, Environmental science & technology.

[46]  R. Hunt,et al.  Approaches to highly parameterized inversion-A guide to using PEST for groundwater-model calibration , 2010 .

[47]  S. P. Neuman.  Universal scaling of hydraulic conductivities and dispersivities in geologic media , 1990 .

[48]  Arlen W. Harbaugh,et al.  MODFLOW-2000, The U.S. Geological Survey Modular Ground-Water Model - User Guide to Modularization Concepts and the Ground-Water Flow Process , 2000 .

[49]  Andres Alcolea,et al.  Inversion of heterogeneous parabolic-type equations using the pilot points method , 2006 .

[50]  Nikolaus Hansen,et al.  Completely Derandomized Self-Adaptation in Evolution Strategies , 2001, Evolutionary Computation.

[51]  S. Sorooshian,et al.  Shuffled complex evolution approach for effective and efficient global minimization , 1993 .

[52]  Alberto Guadagnini,et al.  Nonlocal and localized analyses of conditional mean steady state flow in bounded, randomly nonuniform domains: 1. Theory and computational approach , 1999 .

[53]  Petros Koumoutsakos,et al.  Reducing the Time Complexity of the Derandomized Evolution Strategy with Covariance Matrix Adaptation (CMA-ES) , 2003, Evolutionary Computation.

[54]  J. Gómez-Hernández,et al.  Stochastic conditional inverse modeling of subsurface mass transport: A brief review and the self-calibrating method , 2003 .