Groundwater Level Prediction Research Articles

Combining random forest and graph wavenet for spatial-temporal data prediction

The prosperity of deep learning has revolutionized many machine learning tasks (such as image recognition, natural language processing, etc.). With the widespread use of autonomous sensor networks, the Internet of Things, and crowd sourcing to monitor real-world processes, the volume, diversity, and veracity of spatial-temporal data are expanding rapidly. However, traditional methods have their limitation in coping with spatial-temporal dependencies, which either incorporate too much data from weakly connected locations or ignore the relationships between those interrelated but geographically separated regions. In this paper, a novel deep learning model (termed RF-GWN) is proposed by combining Random Forest (RF) and Graph WaveNet (GWN). In RF-GWN, a new adaptive weight matrix is formulated by combining Variable Importance Measure (VIM) of RF with the long time series feature extraction ability of GWN in order to capture potential spatial dependencies and extract long-term dependencies from the input data. Furthermore, two experiments are conducted on two real-world datasets with the purpose of predicting traffic flow and groundwater level. Baseline models are implemented by Diffusion Convolutional Recurrent Neural Network (DCRNN), Spatial-Temporal GCN (ST-GCN), and GWN to verify the effectiveness of the RF-GWN. The Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Mean Absolute Percentage Error (MAPE) are selected as performance criteria. The results show that the proposed model can better capture the spatial-temporal relationships, the prediction performance on the METR-LA dataset is slightly improved, and the index of the prediction task on the PEMS-BAY dataset is significantly improved. These improvements are extended to the groundwater dataset, which can effectively improve the prediction accuracy. Thus, the applicability and effectiveness of the proposed model RF-GWN in both traffic flow and groundwater level prediction are demonstrated.

Graph neural network for groundwater level forecasting

Accurate prediction of groundwater level (GWL) over a period of time is of great importance for groundwater resources management. Machine learning techniques due to their great performance have recently been used for this problem. Previous methods, however, did not consider the spatial relationships between wells due to the difficulty to handle unstructured well location data. In this paper, a graph neural network (GNN) is used to forecast groundwater dynamics where it can represent each well as a node in the graph. The spatial information is, thus, extracted from an interconnected network using graph convolution layers with a self-adaptive adjacency matrix. The temporal features of the sequence are obtained by gated temporal convolutional networks. The model was applied and evaluated for wells in the southwest area of British Colombia in Canada using data about 11 years (2010–2020). The proposed model performs better in terms of all the defined evaluation metrics, when compared with two baseline models: long short-term memory (LSTM) and gated recurrent units (GRU). Moreover, when the spatial dependencies are completely unknown, the model can still learn them from the data and obtain comparable performance. Furthermore, the proposed model has a high efficiency since it can simultaneously model GWL change for all monitoring wells in the system. We also demonstrated that the spatial dependencies between each well could be intuitively interpreted from the learned adjacency matrix.

How does a combination of numerical modeling, clustering, artificial intelligence, and evolutionary algorithms perform to predict regional groundwater levels?

The prediction of groundwater levels in arid and semi-arid regions is of great importance to tailor the best water management strategies. In this study, we propose a new approach that combines simulation, clustering, and optimization tools for groundwater level prediction. This approach simulates groundwater levels (GWL) using the MODFLOW method, clusters the study aquifer into different clusters using the k-mean method, and predicts regional GWL using the artificial neural network (ANN) and adaptive neuro-fuzzy inference system (ANFIS) methods that were optimized by the Harris Hawks Optimization (HHO), Whale Optimization Algorithm (WOA), and Particle Swarm Optimization (PSO). The efficacy of our approach was evaluated via a case study in northwest Iran. The MODFLOW method simulated the distribution of GWL across the study area with R2 = 0.99, root mean square error (RMSE) = 0.97 m (m) and mean absolute error (MAE) = 0.82 m. The k-means method clustered the aquifer into seven clusters based on the hydraulic conductivity, storage coefficient, groundwater level, groundwater depth, groundwater withdrawal, and aquifer saturation thickness parameters. The prediction of groundwater level for each cluster demonstrated the accurate performance of all optimized models with mean RMSE = 0.6 m and mean absolute percentage error (MAPE) = 0.23 m. The prediction phase identified groundwater level in the previous month (obtained from the MODFLOW method), withdrawal of aquifer, precipitation, temperature, and evaporation as the most influential variables for groundwater levels in different clusters. We recommend the methodology proposed here for the prediction of groundwater levels in different aquifers with heterogeneous characteristics that pose computational burdens and uncertainties.

Comparison of machine learning models for predicting groundwater level, case study: Najafabad region

Comparison of machine learning models for predicting groundwater level, case study: Najafabad region

Groundwater level prediction based on GMS and SVR models under climate change conditions: Case Study—Talesh Plain

Groundwater level prediction based on GMS and SVR models under climate change conditions: Case Study—Talesh Plain

Artificial Neural Network Technique for Groundwater Modelling of Jaspura Block

As a result of economic development and climatic change, arid and semi-arid areas are confronted with significant difficulties in the management of limited freshwater resources. For the most part, groundwater is the most significant water supply in these regions. Groundwater level prediction is a critical component of appropriate sustainable development and must be done correctly. The use of physical-based models is often used in the modelling and prediction of groundwater. However, owing to data shortages in many arid and semi-arid regions, they are not relevant in these areas. The usefulness of data-driven techniques in modelling complicated and nonlinear hydrological processes has been shown in many studies. It is the implementation and comparison of four algorithm-based models for forecasting groundwater levels that is the focal point of this research. This article compares the different Artificial Neural Network method and applies to the Jaspura Block of Banda Districts which is part of the Yamuna River Basin. For the forecast of groundwater levels four distinct algorithms Levenberg Marquardt, Gradient Descent, Scaled Conjugate Gradient and Bayesian regularization are used for optimal design. The ANN training data for input is collected from Recharge and Discharge data while groundwater level data for output layer were being used. The best algorithm comes to the Levenberg Marquardt algorithm in comparison with the other algorithms.

Using machine learning algorithms to predict groundwater levels in Indonesian tropical peatlands

Tropical peatlands play a vital role in the global carbon cycle as large carbon reservoirs and substantial carbon sinks. Indonesia possesses the largest share (65 %) of tropical peat carbon, equal to 57.4 Gt C. Human perturbations such as extensive logging, deforestation and canalization exacerbate water losses, especially during dry seasons, when low precipitation and high evapotranspiration rates combine with the increased drainage to lower groundwater levels. Drying and increasing temperatures of the surface peat exacerbate ignition and wildfire risks within the peat soils. As such, it is critically important to know how groundwater levels in peatlands are changing over space and time. In this study, a multilinear regression model as well as two machine learning algorithms, random forest and extreme gradient boosting, were used to model groundwater level over the study period (2010−12) within a peat dome impacted by drainage canals and multiple wildfires in Central Kalimantan, Indonesia. Although all three models performed well, based on overall fit, spatial modeling of groundwater level results revealed that extreme gradient boosting (R2 = 0.998, RMSE = 0.048 m) outperformed random forest (R2 = 0.997, RMSE = 0.054 m) and multilinear regression (R2 = 0.970, RMSE = 0.221 m) near drainage canals, which are key fire ignition risk locations in the peatlands. Our study also shows that, on average, elevation and precipitation are the most important parameters influencing groundwater level spatiotemporally.

Toward improved lumped groundwater level predictions at catchment scale: Mutual integration of water balance mechanism and deep learning method

Model development in groundwater simulation and physics informed deep learning (DL) has been advancing separately with limited integration. This study develops a general hybrid model for groundwater level (GWL) simulations, wherein water balance-based groundwater processes are embedded as physics constrained recurrent neural layers into prevalent DL architectures. Because of the automatic parameterizing process, physics-informed deep learning algorithm (DLA) equips the hybrid model with enhanced abilities of inferring geological structures of catchment and unobserved groundwater-related processes implicitly. The main purposes of this study are: 1) to explore an optimized data-driven method as alternative to complicated groundwater models; 2) to improve the awareness of hydrological knowledge of DL model for lumped GWL simulation; and 3) to explore the lumped data-driven groundwater models for cross-region applications. The 91 illustrative cases of GWL modeling across the middle eastern continental United States (CONUS) demonstrate that the hybrid model outperforms the pure DL models in terms of prediction accuracy, generality, and robustness. More specifically, the hybrid model outperforms the pure DL models in 78 % of catchments with the improved Δ NSE = 0.129. Meanwhile, the hybrid model simulates more stably with different input strategies. This study reveals the superiority and powerful simulation ability of the DL model with physical constraints, which increases trust in data-driven approaches on groundwater modellings.

Daily Groundwater Level Prediction and Uncertainty Using LSTM Coupled with PMI and Bootstrap Incorporating Teleconnection Patterns Information

Daily groundwater level is an indicator of groundwater resources. Accurate and reliable groundwater level (GWL) prediction is crucial for groundwater resources management and land subsidence risk assessment. In this study, a representative deep learning model, long short-term memory (LSTM), is adopted to predict groundwater level with the selected predictors by partial mutual information (PMI), and bootstrap is employed to generate different samples combination for training many LSTM models, and the predicted values by many LSTM models are used for the uncertainty assessment of groundwater level prediction. Two wells of different climate zones in the USA were used as a case study. Different significant predictors of GWL for two wells were identified by PMI from candidate predictors incorporating teleconnection patterns information. The results show that GWL is significantly affected by antecedent GWL, AO, Niño 3.4, Niño 1 + 2, and precipitation in humid areas, and by antecedent GWL, AO, Niño 3.4, Niño 3, Niño 1 + 2, and PNA in arid areas. Predictor selection can assist in improving the prediction performance of the LSTM model. The relationship between GWL and significant predictors were modeled by the LSTM model, and it achieved higher accuracy in humid areas, while the performance in arid areas was poorer due to limited precipitation information. The performance of LSTM was improved by increasing correlation coefficient (R2) values by 10% and 25% for 2 wells compared to generalized regression neural network (GRNN). Three uncertainty evaluation metrics indicate that LSTM reduced the uncertainty compared to GRNN model. LSTM coupling with PMI and bootstrap can be a promising approach for accurate and reliable groundwater level prediction for different climate zones.

Prediction of future groundwater levels under representative concentration pathway scenarios using an inclusive multiple model coupled with artificial neural networks

Abstract Groundwater (GW) plays a key role in water supply in basins. As global warming and climate change affect groundwater level (GWL), it is important to predict it for planning and managing water resources. This study investigates the GWL of the Yazd-Ardakan Plain basin in Iran for the base period of 1979–2005 and predicts for periods of 2020–2059 and 2060–2099. Lagged temperature and rainfall are used as inputs to hybrid and standalone artificial neural network (ANN) models. In this study, the rat swarm algorithm (RSA), particle swarm optimisation (PSO), salp swarm algorithm (SSA), and genetic algorithm (GA) are used to adjust ANN models. The outcomes of these models are then entered into an inclusive multiple model (IMM) as an ensemble model. In this study, the output of climate models is also inserted into the IMM model to improve the estimation accuracy of temperature, rainfall, and GWL. The monthly average temperature for the base period is 12.9 °C, while average temperatures for 2020–2059 under RCP 4.5 and RCP 8.5 scenarios are 14.5 and 15.1 °C, and for 2060–2099 they are 16.41 and 18.5 °C under the same scenarios, respectively. In future periods, rainfall is low in comparison with the base period. Lagged rainfall and temperature of the base period are inserted into ANN-RSA, ANN-SSA, ANN-PSO, ANN-GA, and ANN models to predict GWL for the base period. Outputs of IMM, ANN, and the five hybrid models (ANN-RSA, ANN-SSA, ANN-PSO, and ANN-GA) indicate that root mean square errors (RMSE) are 2.12, 3.2, 4.58, 6.12, 6.98, and 7.89 m, respectively, in the testing level. It is found that GWL depletion in 2020–2059 under RCP 4.5 and RCP 8.5 scenarios are 0.60–0.88 m and 0.80–1.16 m, and in 2060–2099 under the same scenarios they are 1.49–1.97 m and 1.75–1.98 m, respectively. The results highlight the need to prevent overexploitation of GW in the Ardakan-Yazd Plain to avoid water shortages in the future.

Prediction of Groundwater Level Using Artificial Neural Networks Based on Efficient Input Variables Selection by Partial Mutual Information Algorithm

Prediction of Groundwater Level Using Artificial Neural Networks Based on Efficient Input Variables Selection by Partial Mutual Information Algorithm

Delineating Variabilities of Groundwater Level Prediction Across the Agriculturally Intensive Transboundary Aquifers of South Asia

Groundwater depletion in South Asia’s Himalayan, transboundary Indus-Ganges-Brahmaputra-Meghna (IGBM) rivers basin is among the highest globally. Given the high irrigation demand and population, groundwater sustainability requires an improved understanding of groundwater systems for the accurate prediction of groundwater levels (GWLs). However, the prediction of groundwater system behaviors is a significant challenge since it is dominated by spatiotemporal and subsurface depth-dependent drivers. Earlier studies that address the challenges are mainly based on the short spatial and temporal extent and/or do not separate the renewable (i.e., shallow) vs nonrenewable (i.e., deeper) groundwater signals. Here, we first identified the variable importance of spatial and depth-dependent drivers on GWL in the IGBM basin. Our results indicate a greater influence of anthropogenic factors (i.e., widespread pumping and increased population) in most parts of the IGBM basin, except in the precipitation-dominated basin of the Brahmaputra. Our next purpose was to delineate a multifactorial approach for GWL prediction using the two most used machine learning models (i.e., support vector machine and feed-forward neural network) in the literature. In general, the machine learning model outputs show a good match in comparison to the GWL from the observation wells (n = 2303 distributed across India and Bangladesh) with some limitations in areas with increased groundwater irrigation. We separately compared the results from shallow (<35 m) and deep (>35 m) observation wells, emphasizing the significance of deep groundwater pumping. Our approach highlights the importance of spatiotemporal to multidepth factors in GWL prediction and can be adopted in other parts of the globe to predict GWLs.

Mathematical and Machine Learning Models for Groundwater Level Changes: A Systematic Review and Bibliographic Analysis

With the effects of climate change such as increasing heat, higher rainfall, and more recurrent extreme weather events including storms and floods, a unique approach to studying the effects of climatic elements on groundwater level variations is required. These unique approaches will help people make better decisions. Researchers and stakeholders can attain these goals if they become familiar with current machine learning and mathematical model approaches to predicting groundwater level changes. However, descriptions of machine learning and mathematical model approaches for forecasting groundwater level changes are lacking. This study picked 117 papers from the Scopus scholarly database to address this knowledge gap. In a systematic review, the publications were examined using quantitative and qualitative approaches, and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) was chosen as the reporting format. Machine learning and mathematical model techniques have made significant contributions to predicting groundwater level changes, according to the study. However, the domain is skewed because machine learning has been more popular in recent years, with random forest (RF) methods dominating, followed by the methods of support vector machine (SVM) and artificial neural network (ANN). Machine learning ensembles have also been found to help with aspects of computational complexity, such as performance and training times. Furthermore, compared to mathematical model techniques, machine learning approaches achieve higher accuracies, according to our research. As a result, it is advised that academics employ new machine learning techniques while also considering mathematical model approaches to predicting groundwater level changes.

Groundwater level prediction with machine learning for the Vidisha district, a semi-arid region of Central India

Groundwater is one of the most important freshwater resources, particularly in dry and semi-arid regions with low annual precipitation and frequent droughts. For the planning and the management of water resources, this is very essential to generate a soft computing method that can predict or forecast groundwater table (GWT). The different machine and deep learning algorithms are evaluated in this research work to predict the groundwater table using season, monthly rainfall, and evapotranspiration rate, which are further compared. The performance evaluation of the different algorithms was done using five different statistical parameters, i.e., mean absolute error (MAE), mean squared error (MSE), root-mean-square error (RMSE), R2−score, and Nash-Sutcliffe Efficiency (NSE). From the results of this study, it is concluded that the Multivariate polynomial regression algorithm is the most efficient for predicting the groundwater table with the minimum mean absolute error in the range of 0.63–0.85 and R2-score of 0.82–0.88.

Assessing and forecasting of groundwater level fluctuation in Joypurhat district, northwest Bangladesh, using wavelet analysis and ARIMA modeling

Groundwater resource plays a crucial role for agricultural crop production and socio-economic development in some parts of the world including Bangladesh. Joypurhat district, the northwest part of Bangladesh, a crop production hub, is entirely dependent on groundwater irrigation. A precise assessment and prediction of groundwater level (GWL) can assist long-term groundwater resources (GWR) management, especially in drought-prone agricultural regions. Therefore, this study was carried out to identify trends and magnitude of GWL fluctuation (1980–2019) using the modified Mann–Kendall test, Pettitt’s test, and Sen slope estimators in the drought-prone Joypurhat district, northwest Bangladesh. Time-series data analysis was performed to forecast GWL from 2020 to 2050 using the Auto-Regressive Integrated Moving Average (ARIMA) model. The findings of the MMK test revealed a significant declining trend of GWL, and the trend turning points were identified in the years 1991, 1993, 1997, and 2004, respectively. Results also indicate that the declining rate of GWL varied from 0.104 to 0.159 m/year and the average rate of GWL declination was 0.136 m/year during 1980–2019. The outcomes of wavelet spectrum analysis depicted two significant periods of the declining trend in Khetlal and Akkelpur upazilas. The results obtained from the optimal identified model ARIMA (2, 1, 0) indicate that GWL will decline at a depth of 13.76 m in 2050, and the average declination rate of GWL will be 0.143 m/year in the study area. The predicted results showed a similar declining tendency of GWL from 2020 to 2050, suggesting a disquieting condition, particularly for Khetlal upazila. This research would provide a practical approach for GWL assessment and prediction that could help decision-makers implement long-term GWR management in the study area.

Multi-objective optimization of aquifer storage and recovery operations under uncertainty via machine learning surrogates

Aquifer storage and recovery (ASR) is an important water management approach to store excess surface water into aquifers for later use. Quantitative evaluation of ASR performance is not a trivial task and yet becomes more exacting when uncertainty analysis is added to the dimensionality of the problem. Inclusion of uncertainty into the framework of scheduling optimal ASR operations also increases the level of complexity. This study integrates a surrogate modeling approach coupled with a mixed integer nonlinear programming (MINLP) algorithm to optimize multi-objective ASR operations. The uncertainties are analyzed based upon a thorough sampling of the parameters space as well as a novel analysis of Pareto fronts and variograms of representative solutions. Knee point of representative Pareto fronts is selected for in-depth analysis. As a solution to the dimensionality of the problem, Artificial Neural Network (ANN) is employed to generate surrogate models for predicting groundwater levels and injectate distribution within the aquifer during ASR operations. The computational complexity in building a large number of ANNs and deriving of numerous Pareto fronts via solving the MINLP problem are overcome by the assistance of parallel computing. The results show that optimal ASR operations are highly influenced by hydraulic conductivity and longitudinal dispersivity. Higher hydraulic conductivity values lead to a higher number of active stress periods during storage and recovery phases, which requires large volume of extraction to recover the dispersed injectate. In contrast, higher ratios of longitudinal dispersivity to hydraulic conductivity adversely impact the injectate recovery efficiency. Through meaningful representation of objective function uncertainty by variograms, it is inferred that injectate recovery efficiency is more sensitive to longitudinal dispersivity than hydraulic conductivity.

Prediction of groundwater level using the hybrid model combining wavelet transform and machine learning algorithms

Prediction of groundwater level using the hybrid model combining wavelet transform and machine learning algorithms

Heterogeneous computing with graphical processing unit: improvised back-propagation algorithm for water level prediction

&lt;p&gt;A multitude of research has been rising for predicting the behavior of different real-world problems through machine learning models. An erratic nature occurs due to the augmented behavior and inadequacy of the prerequisite dataset for the prediction of water level over different fundamental models that show flat or low-set accuracy. In this paper, a powerful scaling strategy is proposed for improvised back-propagation algorithm using parallel computing for groundwater level prediction on graphical processing unit (GPU) for the Faridabad region, Haryana, India. This paper aims to propose the new streamlined form of a back-propagation algorithm for heterogeneous computing and to examine the coalescence of artificial neural network (ANN) with GPU for predicting the groundwater level. twenty years of data set from 2001-2020 has been taken into consideration for three input parameters namely, temperature, rainfall, and water level for predicting the groundwater level using parallelized backpropagation algorithm on compute unified device architecture (CUDA). This employs the back-propagation algorithm to be best suited to reinforce learning and performance by providing more accurate and fast results for water level predictions on GPUs as compared to sequential ones on central processing units (CPUs) alone.&lt;/p&gt;

A CNN-LSTM Model Based on a Meta-Learning Algorithm to Predict Groundwater Level in the Middle and Lower Reaches of the Heihe River, China

In this study, a deep learning model is proposed to predict groundwater levels. The model is able to accurately complete the prediction task even when the data utilized are insufficient. The hybrid model that we have developed, CNN-LSTM-ML, uses a combined network structure of convolutional neural networks (CNN) and long short-term memory (LSTM) network to extract the time dependence of groundwater level on meteorological factors, and uses a meta-learning algorithm framework to ensure the network’s performance under sample conditions. The study predicts groundwater levels from 66 observation wells in the middle and lower reaches of the Heihe River in arid regions and compares them with other data-driven models. Experiments show that the CNN-LSTM-ML model outperforms other models in terms of prediction accuracy in both the short term (1 month) and long term (12 months). Under the condition that the training data are reduced by 50%, the MAE of the proposed model is 33.6% lower than that of LSTM. The results of ablation experiments show that CNN-LSTM-ML is 26.5% better than the RMSE of the original CNN-LSTM structure. The model provides an effective method for groundwater level prediction and contributes to the sustainable management of water resources in arid regions.

Application of artificial intelligence models for prediction of groundwater level fluctuations: case study (Tehran-Karaj alluvial aquifer).

The nonlinear groundwater level fluctuations depend on the interaction of many factors such as evapotranspiration, precipitation, groundwater abstraction, and hydrogeological characteristics, making groundwater level prediction a complex task. Groundwater level changes are among the most critical issues in water resource management, which can be predicted to effectively provide management solutions to conserve renewable water resources. Understanding the aquifer status using numerical models is time-consuming and also is associated with inherent uncertainty; therefore, in recent decades, the application of artificial intelligence methods to predict water table fluctuations has significantly gained momentum. In this study, artificial neural network (ANN), fuzzy logic (FL), adaptive neuro-fuzzy inference system (ANFIS), and least square support vector machine (SVM) methods were utilized to predict groundwater level (GWL) with 1-, 2-, and 3-month lead time in Tehran-Karaj plain. Several input scenarios were developed considering groundwater levels, average temperature, total precipitation, total evapotranspiration, and average river flow on a monthly interval. The four error criteria, the correlation coefficient (R), root mean squared error (RMSE), Nash-Sutcliffe efficiency (NSE), and mean absolute error (MAE), were the basis to evaluate the models. Results showed that all the applied methods could provide acceptable GWL prediction, but the ANFIS was the most accurate. However, the ANFIS model showed slightly better performance by yielding R = 0.98 for the training stage and R = 0.98 for the testing stage in the P84 observation well and the second combination of inputs and 1-month lead time. The outcomes also revealed that all the approaches mentioned above could appropriately predict GWL for the leading time of 1 and 2months, but the models provided unsatisfactory results for a 3-month leading time.