Groundwater level prediction using machine learning models: A comprehensive review

Tao, Hai; Hameed, Mohammed Majeed; Marhoon, Haydar Abdulameer; Zounemat‐Kermani, Mohammad; Heddam, Salim; Kim, Sungwon; Sulaiman, Sadeq Oleiwi; Tan, Mou Leong; Sa’adi, Zulfaqar; Mehr, Ali Danandeh; Allawi, Mohammed Falah; Abba, S.I.; Zain, Jasni Mohamad; Falah, Mayadah W.; Jamei, Mehdi; Bokde, Neeraj Dhanraj; Bayatvarkeshi, Maryam; Al-Mukhtar, Mustafa; Bhagat, Suraj Kumar; Tiyasha, Tiyasha; Khedher, Khaled Mohamed; Al‐Ansari, Nadhir; Shahid, Shamsuddin; Yaseen, Zaher Mundher

doi:10.1016/j.neucom.2022.03.014

Cited by 219 publications

(91 citation statements)

References 244 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Their results indicate classical machine learning approaches to be the most widely adopted set of algorithms. In fact, in a number of studies reported by [13] the models demonstrated good performance with relatively small sample sizes, further highlighting the potential of machine learning in groundwater modelling applications.…”

Section: Introductionmentioning

confidence: 91%

“…While still relatively novel, machine learning has gained traction in hydrology and the hydrogeology domain [10][11][12][13]. Studies have demonstrated the use of machine learning for tasks such as groundwater level modelling [14][15][16][17][18], groundwater quality modelling [19,20], groundwater exploration [21], and groundwater storage forecasting [22].…”

Section: Introductionmentioning

confidence: 99%

“…Studies have demonstrated the use of machine learning for tasks such as groundwater level modelling [14][15][16][17][18], groundwater quality modelling [19,20], groundwater exploration [21], and groundwater storage forecasting [22]. Ahmadi and Tao [12,13] provide an in-depth review of the use of the machine learning algorithms for groundwater level modelling. Their results indicate classical machine learning approaches to be the most widely adopted set of algorithms.…”

Section: Introductionmentioning

confidence: 99%

“…The use of machine learning techniques within SADC for groundwater modelling, has seen a slow adoption [12,13]. Several studies, such as [14,[31][32][33], have explored machine learning for groundwater modelling in Southern Africa.…”

Section: Introductionmentioning

confidence: 99%

“…Kombo [33] reported good skill using a hybrid K-Nearest Neighbour-Random Forest (KNN-RF) to predict groundwater level fluctuations in fractured rock aquifers in Kenya, while [31,32] demonstrated reasonable skill in predicting groundwater levels in karts aquifers in South Africa. Nonetheless, additional testing is needed to fully understand the applicability of these models in groundwater modelling applications, and especially within the context of data-scarce aquifers [13].…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

A Comparison of Ensemble and Deep Learning Algorithms to Model Groundwater Levels in a Data-Scarce Aquifer of Southern Africa

et al. 2022

View full text Add to dashboard Cite

Machine learning and deep learning have demonstrated usefulness in modelling various groundwater phenomena. However, these techniques require large amounts of data to develop reliable models. In the Southern African Development Community, groundwater datasets are generally poorly developed. Hence, the question arises as to whether machine learning can be a reliable tool to support groundwater management in the data-scarce environments of Southern Africa. This study tests two machine learning algorithms, a gradient-boosted decision tree (GBDT) and a long short-term memory neural network (LSTM-NN), to model groundwater level (GWL) changes in the Shire Valley Alluvial Aquifer. Using data from two boreholes, Ngabu (sample size = 96) and Nsanje (sample size = 45), we model two predictive scenarios: (I) predicting the change in the current month’s groundwater level, and (II) predicting the change in the following month’s groundwater level. For the Ngabu borehole, GBDT achieved R2 scores of 0.19 and 0.14, while LSTM achieved R2 scores of 0.30 and 0.30, in experiments I and II, respectively. For the Nsanje borehole, GBDT achieved R2 of −0.04 and −0.21, while LSTM achieved R2 scores of 0.03 and −0.15, in experiments I and II, respectively. The results illustrate that LSTM performs better than the GBDT model, especially regarding slightly greater time series and extreme GWL changes. However, closer inspection reveals that where datasets are relatively small (e.g., Nsanje), the GBDT model may be more efficient, considering the cost required to tune, train, and test the LSTM model. Assessing the full spectrum of results, we concluded that these small sample sizes might not be sufficient to develop generalised and reliable machine learning models.

show abstract

Section: Introductionmentioning

confidence: 91%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

A Comparison of Ensemble and Deep Learning Algorithms to Model Groundwater Levels in a Data-Scarce Aquifer of Southern Africa

et al. 2022

View full text Add to dashboard Cite

show abstract

Sustainable Management of Natural Resources for Environmental Sustainability

Ismail

Pardi

Radzun

et al. 2023

Ecorestoration for Sustainability

View full text Add to dashboard Cite

Large-scale groundwater monitoring in Brazil assisted with satellite-based artificial intelligence techniques

Camacho

Getirana²,

Filho

et al. 2022

Preprint

View full text Add to dashboard Cite

Here, we develop and test an artificial intelligence (AI)-based approach to monitor major Brazilian aquifers. The approach combines Gravity Recovery and Climate Experiment (GRACE) data and ground-based hydrogeological measurements from Brazil's Integrated Groundwater Monitoring Network at hundreds of wells distributed in twelve aquifers across the country. We use a model ensemble composed of four different AI models: Extreme Gradient Boost, Light Gradient Boosting Model, CatBoost and Multilayer Perceptron. The approach is further boosted with wavelet and seasonal decomposition processes applied to GRACE data. To determine the sensitivity of the AI approach to data availability, we propose four experiments combining hydrogeological measurements from different aquifers. Groundwater storage estimates from the Global Land Data Assimilation System (GLDAS) are used as the benchmark. The AI approach successfully reproduces groundwater storage estimates at all Brazilian aquifers. Results show that the proposed approach outperforms GLDAS in all experiments, with an average Nash-Sutcliff efficiency of 0.91 and an average RMSE of 0.43cm for the experiment that covers all monitored wells in Brazil. GLDAS resulted in -1.311 and 5.84cm, respectively. This study demonstrates that combining satellite data and AI can be a cost-effective alternative to monitor poorly equipped aquifers at the continental scale.

show abstract

Groundwater level prediction using machine learning models: A comprehensive review

Cited by 219 publications

References 244 publications

A Comparison of Ensemble and Deep Learning Algorithms to Model Groundwater Levels in a Data-Scarce Aquifer of Southern Africa

A Comparison of Ensemble and Deep Learning Algorithms to Model Groundwater Levels in a Data-Scarce Aquifer of Southern Africa

Sustainable Management of Natural Resources for Environmental Sustainability

Large-scale groundwater monitoring in Brazil assisted with satellite-based artificial intelligence techniques

Contact Info

Product

Resources

About