Elements of complexity in subsurface modeling, exemplified with three case studies

There are complexity elements to consider when applying subsurface flow and transport models to support environmental analyses. Modelers balance the benefits and costs of modeling along the spectrum of complexity, taking into account the attributes of more simple models (e.g., lower cost, faster execution, easier to explain, less mechanistic) and the attributes of more complex models (higher cost, slower execution, harder to explain, more mechanistic and technically defensible). In this report, modeling complexity is examined with respect to considering this balance. The discussion of modeling complexity is organized into three primary elements: (1) modeling approach, (2) description of process, and (3) description of heterogeneity. Three examples are used to examine these complexity elements. Two of the examples use simulations generated from a complex model to develop simpler models for efficient use in model applications. The first example is designed to support performance evaluation of soil-vapor-extraction remediation in terms of groundwater protection. The second example investigates the importance of simulating different categories of geochemical reactions for carbon sequestration and selecting appropriate simplifications for use in evaluating sequestration scenarios. In the third example, the modeling history for a uranium-contaminated site demonstrates that conservative parameter estimates were inadequate surrogates for complex, critical processes and there is discussion on the selection of more appropriate model complexity for this application. All three examples highlight how complexity considerations are essential to create scientifically defensible models that achieve a balance between model simplification and complexity.RésuméIl y a des éléments de complexité à prendre en considération lorsqu’on applique des modèles d’écoulements et de transport souterrains dans le cadre d’analyses environnementales. Les modélisateurs recherchent un équilibre entre les bénéfices et les coûts de la modélisation selon le spectre de la complexité, prenant en considération les caractéristiques de modèles plus simples (c’est-à-dire moins coûteux, exécution plus rapide, plus faciles à expliquer, moins mécanistes) et des caractéristiques des modèles plus complexes (c’est-à-dire plus coûteux, exécution plus lente, plus difficiles à expliquer, plus mécanistes et techniquement fiables). Dans cet article, la complexité en modélisation est étudiée afin d’atteindre cet équilibre. La discussion de la complexité en modélisation s’articule autour de trois principaux éléments : (1) approche de modélisation, (2) description des processus, et (3) description des hétérogénéités. Trois exemples sont utilisés pour étudier ces éléments de complexité. Deux de ces exemples utilisent des simulations résultant d’un modèle complexe afin de développer des modèles plus simples pour une utilisation efficace dans l’application du modèle. Le premier exemple est conçu pour évaluer la performance d’une remédiation par extraction de vapeur du sol pour la protection des eaux souterraines. Le deuxième exemple examine l’importance de différents types de simulation de réactions géochimiques pour le stockage du CO2, en sélectionnant des simplifications appropriées pour évaluer des scénarios de stockage. Dans le troisième exemple, l’histoire de la modélisation pour un site pollué à l’uranium démontre que les estimations des paramètres conservatifs étaient des substitutifs inadéquats pour des processus complexes et critiques et la sélection d’une complexité de modèle plus adaptée est discutée pour cette application. Ces trois exemples mettent en évidence comment la prise en considération de la complexité est essentielle pour réaliser des modèles scientifiquement fiables qui atteignent un équilibre entre simplification et complexité du modèle.ResumenHay elementos de complejidad a considerar cuando se aplican modelos de flujo y transporte en el subsuelo para apoyar los análisis ambientales. Los modelistas equilibran los beneficios y costos del modelado a lo largo del espectro de complejidad, teniendo en cuenta los atributos de los modelos más simples (por ejemplo, menor costo, ejecución más rápida, más fácil de explicar, menos mecanicista) y los atributos de modelos más complejos (costo, ejecución más lenta, más difícil de explicar, más mecanicista y técnicamente defendible). En este trabajo, se examina la complejidad del modelado con respecto a considerar este balance. La discusión de la complejidad de modelado se organiza en tres elementos principales: (1) enfoque del modelado, (2) descripción del proceso y (3) descripción de la heterogeneidad. Se utilizan tres ejemplos para examinar estos elementos de complejidad. Dos de los ejemplos utilizan simulaciones generadas a partir de un modelo complejo para desarrollar modelos más simples para un uso eficiente en aplicaciones de modelos. El primer ejemplo está diseñado para apoyar la evaluación del rendimiento de la remediación para la extracción de vapores del suelo en términos de protección del agua subterránea. El segundo ejemplo investiga la importancia de simular diferentes categorías de reacciones geoquímicas para el secuestro de carbono y seleccionar las simplificaciones apropiadas para su uso en la evaluación de escenarios de secuestro. En el tercer ejemplo, el historial de modelado de un sitio contaminado con uranio demuestra que las estimaciones de parámetros conservadores eran substitutos inadecuados para procesos críticos complejos y hay una discusión sobre la selección de la complejidad de modelo más apropiada para esta aplicación. Los tres ejemplos ponen de relieve cómo las consideraciones de complejidad son esenciales para crear modelos científicamente defendibles que logren un balance entre la simplificación y la complejidad del modelo.摘要应用地表以下水流和运移模型支持环境分析时,有一些复杂性元素需要考虑。建模者在复杂性范围内平衡着建模的效益和成本,要考虑较简单模型的属性(较低成本、较快的实施、易于说明及不那么机械)及较复杂模型的属性(较高的成本、较慢的实施、很难说明及技术上可防御)。在本文中,考虑到这个平衡问题,检查了建模的复杂性。建模复杂性的论述被归纳为三个主要元素:(1)建模方法,(2)过程描述,(3)异质性描述。利用三个例子检查了复杂性元素。为了在模型应用中有效利用模型,从一个复杂模型开发出较简单的模型,其中两个例子使用了所产生的模拟结果。第一个例子被设计为支持地下水保护方面土壤气体萃取修复性能评价。第二个例子调查了针对碳隔离模拟不同类别地球化学反应及选择在隔离方案中使用适当简单化的重要性。在第三个例子中,一个铀污染场地的建模历史说明,保守参数估算不能适当地值代替复杂的、关键过程,应用这个模型在更恰当模型复杂性的选择上有争论。所有三个例子强调了在科学上创建防御性模型取得模型简单化和复杂性之间平衡上,复杂性考量是多么的重要。ResumoExistem elementos de complexidade a se considerar quando se aplica modelos de transporte e fluxo de subsuperficie para auxiliar análises ambientais. Os modeladores ponderam os benefícios e custos da modelagem ao longo do espectro da complexidade, levando em consideração os atributos de modelos mais simples (p.ex. menor custo, execução mais rápida, mais fácil de explicar, menos mecanicista) e os atributos de modelos mais complexos (maior custo, execução mais lenta, mais difícil de explicar, mais mecanicista e tecnicamente defensável). Nesse estudo, a complexidade da modelagem é examinada considerando esse balanço. A discussão da complexidade da modelagem está organizada em três elementos principais: (1) abordagem da modelagem, (2) descrição do processo, e (3) descrição da heterogeneidade. Três exemplos são utilizados para examinar esses elementos de complexidade. Dois dos exemplos utilizam simulações geradas através de um modelo complexo para desenvolver modelos mais simples, para o uso eficiente em aplicações de modelos. O primeiro exemplo é projetado para auxiliar a avaliação do desempenho da remediação por extração de vapor do solo em termos de proteção das águas subterrâneas. O segundo exemplo investiga a importância de simular diferentes categorias de reações geoquímicas para sequestro de carbono e selecionar as simplificações apropriadas para uso na avaliação de cenários de sequestro. No terceiro exemplo, o histórico de modelagem de uma área contaminada com urânio demonstra que as estimativas de parâmetros conservadores foram substitutos inadequados para processos críticos e complexos e há uma discussão sobre a seleção de um modelo com complexidade mais apropriada para essa aplicação. Todos os três exemplos destacam como considerações de complexidade são essenciais para criar modelos cientificamente defensáveis que alcancem um equilíbrio entre simplificação e complexidade do modelo.

[1]  Mark L Brusseau,et al.  Assessing performance and closure for soil vapor extraction: integrating vapor discharge and impact to groundwater quality. , 2012, Journal of contaminant hydrology.

[2]  Craig M. Bethke,et al.  How the Kd Approach Undermines Ground Water Cleanup , 2000 .

[3]  Eugene J. Freeman,et al.  Contaminants of Potential Concern in the 300-FF-5 Operable Unit: Expanded Annual Groundwater Report for Fiscal Year 2004 , 2005 .

[5]  G. Nützmann,et al.  Model based verification and prognosis of acidification and sulphate releasing processes downstream of a former sewage field in Berlin (Germany). , 2009, Journal of contaminant hydrology.

[6]  G. V. Last,et al.  Evaluation of deep vadose zone contaminant flux into groundwater: Approach and case study. , 2016, Journal of contaminant hydrology.

[7]  M. Oostrom,et al.  Soil Vapor Extraction System Optimization, Transition, and Closure Guidance , 2013 .

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

[9]  Christopher F. Brown,et al.  Key factors for determining groundwater impacts due to leakage from geologic carbon sequestration reservoirs , 2014 .

[10]  J. Doherty Modeling: Picture Perfect or Abstract Art? , 2011, Ground water.

[11]  A. Dassargues,et al.  Application of multiple-point geostatistics on modelling groundwater flow and transport in a cross-bedded aquifer (Belgium) , 2009 .

[12]  K. Carroll,et al.  Estimating the Impact of Vadose Zone Sources on Groundwater to Support Performance Assessment of Soil Vapor Extraction , 2014 .

[13]  Steven B. Yabusaki,et al.  Flow and Transport in the Hanford 300 Area Vadose Zone-Aquifer-River System , 2005 .

[14]  George Shu Heng Pau,et al.  A high-performance workflow system for subsurface simulation , 2014, Environ. Model. Softw..

[15]  Yilin Fang,et al.  Variably saturated flow and multicomponent biogeochemical reactive transport modeling of a uranium bioremediation field experiment. , 2011, Journal of contaminant hydrology.

[16]  Mary C Hill,et al.  The Practical Use of Simplicity in Developing Ground Water Models , 2006, Ground water.

[17]  R. Liedl,et al.  Complexity vs. Simplicity: Groundwater Model Ranking Using Information Criteria , 2014, Ground water.

[18]  Paul D. Thorne,et al.  Carbon Tetrachloride Flow and Transport in the Subsurface of the 216‐Z‐9 Trench at the Hanford Site , 2007 .

[19]  Henning Prommer,et al.  A field‐scale reactive transport model for U(VI) migration influenced by coupled multirate mass transfer and surface complexation reactions , 2010 .

[20]  M. Wheeler,et al.  A Two‐Tiered Approach to Reactive Transport: Application to Sr Mobility Under Variable pH , 1998 .

[21]  S. Bhattacharjee,et al.  Virus transport in physically and geochemically heterogeneous subsurface porous media. , 2002, Journal of contaminant hydrology.

[22]  Glenn E. Hammond,et al.  Field‐scale model for the natural attenuation of uranium at the Hanford 300 Area using high‐performance computing , 2010 .

[23]  Philip E. Long,et al.  Microbiological and Geochemical Heterogeneity in an In Situ Uranium Bioremediation Field Site , 2005, Applied and Environmental Microbiology.

[24]  Hajime Yamamoto,et al.  Investigation of gridding effects for numerical simulations of CO2 geologic sequestration , 2011 .

[25]  E. Reardon,et al.  Kd's — Can They Be Used to Describe Reversible Ion Sorption Reactions in Contaminant Migration? , 1981 .

[26]  P. Lichtner,et al.  Stochastic simulation of uranium migration at the Hanford 300 Area. , 2011, Journal of contaminant hydrology.

[27]  Gour-Tsyh Yeh,et al.  A Model for Simulating Transport of Reactive Multispecies Components: Model Development and Demonstration , 1991 .

[28]  Threshold Values for Identification of Contamination Predicted by Reduced-Order Models , 2014 .

[29]  John M. Zachara,et al.  Scale‐dependent desorption of uranium from contaminated subsurface sediments , 2008 .

[30]  R. L. Naff,et al.  Nonreactive and reactive solute transport in three-dimensional heterogeneous porous media: Mean displacement, plume spreading, and uncertainty , 1994 .

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

[32]  S. Heald,et al.  Dissolution of uranyl microprecipitates in subsurface sediments at Hanford Site, USA , 2004 .

[33]  Rajesh J. Pawar,et al.  Reduced order models for assessing CO 2 impacts in shallow unconfined aquifers , 2016 .

[34]  Kyle R. Parker,et al.  System-Scale Model of Aquifer, Vadose Zone, and River Interactions for the Hanford 300 Area - Application to Uranium Reactive Transport , 2013 .

[35]  G. Zyvoloski,et al.  Grid Resolution Study of Ground Water Flow and Transport , 2005, Ground water.

[36]  Christopher F. Brown,et al.  A Site Wide Perspective on Uranium Geochemistry at the Hanford Site , 2007 .

[37]  Keith Beven,et al.  Prophecy, reality and uncertainty in distributed hydrological modelling , 1993 .

[38]  Jens T. Birkholzer,et al.  On modeling the potential impacts of CO2 sequestration on shallow groundwater: Transport of organics and co-injected H2S by supercritical CO2 to shallow aquifers , 2013 .

[39]  A. Ward,et al.  Establishing a geochemical heterogeneity model for a contaminated vadose zone--aquifer system. , 2013, Journal of contaminant hydrology.

[40]  George A Zyvoloski,et al.  An Investigation of Numerical Grid Effects in Parameter Estimation , 2006, Ground water.

[41]  Hari S. Viswanathan,et al.  Significance of carbonate buffers in natural waters reacting with supercritical CO2: Implications for monitoring, measuring and verification (MMV) of geologic carbon sequestration , 2008 .

[42]  Simulating adsorption of U(VI) under transient groundwater flow and hydrochemistry: Physical versus chemical nonequilibrium model , 2011 .

[43]  Randall J. Hunt,et al.  Debating complexity in modeling , 1999 .

[44]  D. Freyberg,et al.  On using the equivalent conductivity to characterize solute spreading in environments with low‐permeability lenses , 2002 .

[45]  M. Oostrom,et al.  Sensor and Numerical Simulator Evaluation for Porous Medium Desiccation and Rewetting at the Intermediate Laboratory Scale , 2012 .

[46]  J. H. Dane,et al.  Desiccation of Unsaturated Porous Media: Intermediate‐Scale Experiments and Numerical Simulation , 2009 .

[47]  Zhenxue Dai,et al.  Geochemical impacts of carbon dioxide, brine, trace metal and organic leakage into an unconfined, oxidizing limestone aquifer , 2014 .

[48]  Michael J. Truex,et al.  Estimating Persistent Mass Flux of Volatile Contaminants from the Vadose Zone to Ground Water , 2009 .

[49]  Mark D. Freshley,et al.  Updated Conceptual Model for the 300 Area Uranium Groundwater Plume , 2012 .

[50]  Mark D. Freshley,et al.  Scientific Opportunities for Monitoring at Environmental Remediation Sites (SOMERS): Integrated Systems-Based Approaches to Monitoring , 2012 .

[51]  Keith Beven,et al.  Equifinality, data assimilation, and uncertainty estimation in mechanistic modelling of complex environmental systems using the GLUE methodology , 2001 .

[52]  Ming Ye,et al.  Combined Estimation of Hydrogeologic Conceptual Model, Parameter, and Scenario Uncertainty with Application to Uranium Transport at the Hanford Site 300 Area , 2006 .

[53]  S. Benson,et al.  Accurate determination of characteristic relative permeability curves , 2015 .

[54]  M. Celia,et al.  Effects of heterogeneities on capillary pressure-saturation-relative permeability relationships. , 2002, Journal of contaminant hydrology.

[55]  T Prabhakar Clement,et al.  Complexities in Hindcasting Models—When Should We Say Enough Is Enough? , 2011, Ground water.

[56]  Steven M. Gorelick,et al.  Geological modeling of submeter scale heterogeneity and its influence on tracer transport in a fluvial aquifer , 2010 .

[57]  Michael J. Truex,et al.  Three‐Dimensional Simulation of Volatile Organic Compound Mass Flux from the Vadose Zone to Groundwater , 2010 .

[58]  Christopher F. Brown,et al.  Source term modeling for evaluating the potential impacts to groundwater of fluids escaping from a depleted oil reservoir used for carbon sequestration , 2014 .

[59]  Shaoping Chu,et al.  The National Risk Assessment Partnership’s integrated assessment model for carbon storage: A tool to support decision making amidst uncertainty , 2016 .

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

[61]  D. Kent,et al.  Quantifying Differences in the Impact of Variable Chemistry on Equilibrium Uranium(VI) Adsorption Properties of Aquifer Sediments , 2011, Environmental science & technology.

[62]  Yousu Chen,et al.  Three-Dimensional Groundwater Models of the 300 Area at the Hanford Site, Washington State , 2008 .

[63]  R. Martel,et al.  Accounting for Aquifer Heterogeneity from Geological Data to Management Tools , 2012, Ground water.

[64]  R. Hunt,et al.  Are Models Too Simple? Arguments for Increased Parameterization , 2007, Ground water.

[65]  J. Zachara,et al.  Kinetics of uranium(VI) desorption from contaminated sediments: effect of geochemical conditions and model evaluation. , 2009, Environmental science & technology.

[66]  F. Felletti,et al.  Hierarchical simulation of aquifer heterogeneity: implications of different simulation settings on solute-transport modeling , 2016, Hydrogeology Journal.