Hydrogeophysical Characterization of Nonstationary DNAPL Source Zones by Integrating a Convolutional Variational Autoencoder and Ensemble Smoother

Detailed characterization of dense nonaqueous phase liquid (DNAPL) source zone architecture (SZA) is essential for designing efficient remediation strategies. However, it is difficult to characterize a highly irregular and localized SZA, because traditional drilling investigations provide limited in...

Full description

Saved in:
Bibliographic Details
Published in:Water resources research Vol. 57; no. 2
Main Authors: Kang, Xueyuan, Kokkinaki, Amalia, Kitanidis, Peter K., Shi, Xiaoqing, Lee, Jonghyun, Mo, Shaoxing, Wu, Jichun
Format: Journal Article
Language:English
Published: Washington John Wiley & Sons, Inc 01.02.2021
Subjects:
Aquifers
Contaminants
convolutional variational autoencoder
Data
data collection
Datasets
Deep learning
deep‐learning parameterization
Depletion
DNAPL source zone characterization
Drilling
ensemble smoother
Estimation accuracy
Frameworks
Geology
Geophysical data
geophysics
geostatistics
Hydrogeology
Identification methods
Infiltration
Iterative methods
joint hydrogeophysical inversion
Learning
Learning behaviour
Liquids
Mathematical models
Modelling
Morphology
Multiphase
Nonaqueous phase liquids
Numerical experiments
Organic contaminants
Parameterization
Performance evaluation
Physics
Remediation
Saturation
Statistical methods
Statistical models
Training
water
ISSN:0043-1397, 1944-7973
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Detailed characterization of dense nonaqueous phase liquid (DNAPL) source zone architecture (SZA) is essential for designing efficient remediation strategies. However, it is difficult to characterize a highly irregular and localized SZA, because traditional drilling investigations provide limited information. With limited data, the estimation accuracy of traditional geostatistical methods is strongly affected by the parameterization of the prior description of the SZA. To improve characterization performance, we parameterized the DNAPL saturation field using a physics‐based approach. We trained a convolutional variational autoencoder (CVAE) using data from multiphase modeling that captures the physics of DNAPL infiltration. The trained CVAE network was used in SZA inversion to obtain an improved prior DNAPL saturation field, instead of the typical stationary prior covariances. We then integrated the CVAE network into an iterative ensemble smoother (ES), to formulate a joint inversion framework. To overcome difficulties from limited/sparse data, we incorporated hydrogeological and geophysical datasets in the proposed inversion framework. To evaluate the performance of our method, we conducted numerical experiments in a hypothetical heterogeneous aquifer with an intricate SZA. The results show that the CVAE was an effective and efficient parameterization method which can capture the DNAPL infiltration patterns better than a Gaussian prior. The improved prior, combined with multisource datasets, can result in better resolution, and overall improved SZA characterization. In contrast to the standard ES method, the proposed framework reconstructed the SZA more accurately. We also demonstrated that DNAPL depletion behavior and dissolved concentration profiles can be predicted accurately using the estimated SZA. Plain Language Summary Toxic organic contaminants, such as chlorinated solvents, often exist in the form of oily phases that do not mix with water and are denser than water. Such dense nonaqueous phase liquids (DNAPLs) represent a challenging environmental problem worldwide. To develop efficient remediation strategies, the quantity and morphology of the DNAPL in the subsurface needs to be identified. However, state‐of‐the‐art identification methods cannot reproduce accurate and physically realistic patterns of DNAPL source zone architecture (SZA). In this work, we developed a deep‐learning strategy that can satisfactorily capture the physical patterns of the DNAPL SZA by learning these patterns from training samples obtained from multiphase modeling. This parameterization strategy was further integrated with an ensemble‐based inversion method to estimate the SZA by incorporating multisource data sets. The proposed framework produced more accurate estimates of both the quantity and distribution of DNAPL than the results from traditional inversion methods that utilized a purely statistical model to represent the SZA when using data to finetune the characterization. Key Points A variational autoencoder (VAE) is proposed for parameterizing complex dense nonaqueous phase liquid (DNAPL) source zones, preserving their nonstationary spatial characteristics Integrating VAE and ensemble smoother (ES) outperforms the standard ES in identifying both the residual and pooled DNAPL Combining hydrogeological and geophysical data sets can improve the characterization of the nonstationary DNAPL distribution
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
content type line 23
ISSN:0043-1397
1944-7973
DOI:10.1029/2020WR028538