A Deep Learning-Based Data Assimilation Approach to Characterizing Coastal Aquifers Amid Non-linearity and Non-Gaussianity Challenges DOI Open Access
Chenglong Cao, Jiangjiang Zhang,

Gan Wei

и другие.

Authorea (Authorea), Год журнала: 2024, Номер unknown

Опубликована: Июнь 3, 2024

Seawater intrusion (SI) poses a substantial threat to water security in coastal regions, where numerical models play pivotal role supporting groundwater management and protection. However, the inherent heterogeneity of aquifers introduces significant uncertainties into SI predictions, potentially diminishing their effectiveness decisions. Data assimilation (DA) offers solution by integrating various types observational data with model characterize heterogeneous aquifers. Traditional DA techniques, like ensemble smoother using Kalman formula (ES) Markov chain Monte Carlo, face challenges when confronted non-linearity, non-Gaussianity, high-dimensionality issues commonly encountered aquifer characterization. In this study, we introduce novel approach rooted deep learning (DL), referred as ES, aimed at effectively characterizing varying levels heterogeneity. We systematically investigate range factors that impact performance including number observations, degree heterogeneity, structure training options DL models. Our findings reveal ES excels under non-linear non-Gaussian conditions. Comparison between different experimentation settings underscores robustness ES. Conversely, certain scenarios, displays noticeable biases characterization results, especially measurement from discontinuous processes are used. To optimize efficacy attention must be given design selection data, which crucial ensure universal applicability method.

Язык: Английский

Parameter estimation and uncertainty quantification of rainfall-runoff models using data assimilation methods based on deep learning and local ensemble updates DOI

Lei Yao,

Jiangjiang Zhang, Chenglong Cao

и другие.

Environmental Modelling & Software, Год журнала: 2025, Номер unknown, С. 106332 - 106332

Опубликована: Янв. 1, 2025

Язык: Английский

Процитировано

2
A Deep Learning‐Based Data Assimilation Approach to Characterizing Coastal Aquifers Amid Non‐Linearity and Non‐Gaussianity Challenges DOI Creative Commons
Chenglong Cao, Jiangjiang Zhang,

Gan Wei

и другие.

Water Resources Research, Год журнала: 2024, Номер 60(7)

Опубликована: Июнь 30, 2024

Abstract Seawater intrusion (SI) poses a substantial threat to water security in coastal regions, where numerical models play pivotal role supporting groundwater management and protection. However, the inherent heterogeneity of aquifers introduces significant uncertainties into SI predictions, potentially diminishing their effectiveness decisions. Data assimilation (DA) offers solution by integrating various types observational data with model characterize heterogeneous aquifers. Traditional DA techniques, like ensemble smoother using Kalman formula (ES K ) Markov chain Monte Carlo, face challenges when confronted non‐linearity, non‐Gaussianity, high‐dimensionality issues commonly encountered aquifer characterization. In this study, we introduce novel approach rooted deep learning (DL), referred as ES DL , aimed at effectively characterizing varying levels heterogeneity. We systematically investigate range factors that impact performance including number observations, degree heterogeneity, structure training options models. Our findings reveal excels under non‐linear non‐Gaussian conditions. Comparison between different experimentation settings underscores robustness . Conversely, certain scenarios, displays noticeable biases characterization results, especially measurement from discontinuous processes are used. To optimize efficacy attention must be given design selection data, which crucial ensure universal applicability method.

Язык: Английский

Процитировано

3
AquaCrop model-based sensitivity analysis of soil salinity dynamics and productivity under climate change in sandy-layered farmland DOI Creative Commons
Zhuangzhuang Feng, Qingfeng Miao, Haibin Shi

и другие.

Agricultural Water Management, Год журнала: 2024, Номер 307, С. 109244 - 109244

Опубликована: Дек. 25, 2024

Язык: Английский

Процитировано

1
A Deep Learning-Based Data Assimilation Approach to Characterizing Coastal Aquifers Amid Non-linearity and Non-Gaussianity Challenges DOI Open Access
Chenglong Cao, Jiangjiang Zhang,

Gan Wei

и другие.

Authorea (Authorea), Год журнала: 2024, Номер unknown

Опубликована: Июнь 3, 2024

Seawater intrusion (SI) poses a substantial threat to water security in coastal regions, where numerical models play pivotal role supporting groundwater management and protection. However, the inherent heterogeneity of aquifers introduces significant uncertainties into SI predictions, potentially diminishing their effectiveness decisions. Data assimilation (DA) offers solution by integrating various types observational data with model characterize heterogeneous aquifers. Traditional DA techniques, like ensemble smoother using Kalman formula (ES) Markov chain Monte Carlo, face challenges when confronted non-linearity, non-Gaussianity, high-dimensionality issues commonly encountered aquifer characterization. In this study, we introduce novel approach rooted deep learning (DL), referred as ES, aimed at effectively characterizing varying levels heterogeneity. We systematically investigate range factors that impact performance including number observations, degree heterogeneity, structure training options DL models. Our findings reveal ES excels under non-linear non-Gaussian conditions. Comparison between different experimentation settings underscores robustness ES. Conversely, certain scenarios, displays noticeable biases characterization results, especially measurement from discontinuous processes are used. To optimize efficacy attention must be given design selection data, which crucial ensure universal applicability method.

Язык: Английский

Процитировано

0