Groundwater level reconstruction using long-term climate reanalysis data and deep neural networks

Chidepudi, Sivarama Krishna Reddy; Massei, Nicolas; Jardani, Abderrahim; Henriot, Abel

doi:10.1016/j.ejrh.2023.101632

Cited by 3 publications

(3 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Wavelet pre-processing generally improves model performance, especially in the inertial GWL category, where cumulative distribution functions (CDFs) are steeper and shifted to the right, indicating a higher proportion of simulations with high performance. This is in line with previous findings as already reported in our previous works 315 (Chidepudi et al, 2023a(Chidepudi et al, & 2024. This demonstrates the wavelet decomposition ability to extract "hidden" inertial dynamics features which facilitates their assimilation by the model in the learning process.…”

supporting

confidence: 93%

“…CC BY 4.0 License. applied to hydrological simulations, as detailed in several studies (Chidepudi et al, 2023a;Chidepudi et al, 2024;Fang et al, 2022;Kratzert et al, 2021;Li et al, 2022;Vu et al, 2023). To further interpret and decrypt the results for better understanding, we used the SHAP approach (Lundberg & Lee, 2017), which is an increasingly popular game-centric approach for 230 explaining the outcomes of deep learning models.…”

Section: Datamentioning

confidence: 99%

See 1 more Smart Citation

Comment on egusphere-2024-794

2024

View full text Add to dashboard Cite

supporting

confidence: 93%

Section: Datamentioning

confidence: 99%

Comment on egusphere-2024-794

2024

View full text Add to dashboard Cite

Section: Datamentioning

confidence: 99%

Reply on RC1

Chidepudi

2024

View full text Add to dashboard Cite

Training deep learning models with a multi-station approach and static aquifer attributes for groundwater level simulation: what’s the best way to leverage regionalised information?

Chidepudi,

Massei,

Jardani

et al. 2024

Preprint

View full text Add to dashboard Cite

Abstract. In this study, we used deep learning models with recurrent structure neural networks to simulate large-scale groundwater level (GWL) fluctuations in northern France. We developed a multi-station collective training for GWL simulations, using both “dynamic” variables (i.e. climatic) and static aquifer characteristics. This large-scale approach offers the possibility of incorporating dynamic and static features to cover more reservoir heterogeneities in the study area. Further, we investigated the performance of relevant feature extraction techniques such as clustering and wavelet transform decomposition, intending to simplify network learning using regionalised information. Several modelling performance tests were conducted. Models specifically trained on different types of GWL, clustered based on the spectral properties of the data, performed significantly better than models trained on the whole dataset. Clustering-based modelling reduces complexity in the training data and targets relevant information more efficiently. Applying multi-station models without prior clustering can lead the models to learn the dominant station behavior preferentially, ignoring unique local variations. In this respect, wavelet pre-processing was found to partially compensate clustering, bringing out common temporal and spectral characteristics shared by all available time series even when these characteristics are “hidden” because of too small amplitude. When employed along with prior clustering, thanks to its capability of capturing essential features across all time scales (high and low), wavelet decomposition used as a pre-processing technique provided significant improvement in model performance, particularly for GWLs dominated by low-frequency variations. This study advances our understanding of GWL simulation using deep learning, highlighting the importance of different model training approaches, the potential of wavelet preprocessing, and the value of incorporating static attributes.

show abstract

Groundwater level reconstruction using long-term climate reanalysis data and deep neural networks

Cited by 3 publications

References 27 publications

Comment on egusphere-2024-794

Comment on egusphere-2024-794

Reply on RC1

Training deep learning models with a multi-station approach and static aquifer attributes for groundwater level simulation: what’s the best way to leverage regionalised information?

Contact Info

Product

Resources

About