Groundwater level forecasting using artificial neural networks

Hydrological Sciences Journal, 2013

Hydrological Sciences Journal/Journal des Sciences Hydrologiques

Artificial Neural Networks (ANNs) have been lately used to predict the hydraulic head in well locations. In the present work, Particle Swarm Optimization (PSO) algorithm is used to train a feed-forward multi-layer ANN for the simulation of hydraulic head change at an observation well at the region of Agia, Chania, Greece. Three variants of the particle swarm optimization algorithm are considered, the classic one with the inertia weight improvement, PSO-TVAC and GLBest-PSO. The best performance was achieved by GLBest-PSO when implemented using field data from the region of interest, providing improved training results compared to the back propagation training algorithm. The trained ANN was subsequently used for midterm prediction of the hydraulic head as well as for the study of three climate change scenarios. Data time series were created using a stochastic weather generator, and the scenarios were examined for the period 2010-2020.

Journal of Water and Wastewater, 2021

Groundwater is an important source of freshwater the world over, especially in arid and semiarid regions. In recent years, groundwater overextraction has led to a serious drawdown in groundwater level in many aquifers. Hence, the projecting groundwater level is essential for the planning and management of water resources in a basin scale. This study aimed to project the mean groundwater level in Najafabad Plain in central Iran. Najafabad Plain is one of the most important aquifers in the Zayandeh-Rud River basin currently facing a negative hydrologic balance, which has been aggravated by the excessive agricultural demand that has adversely affected its groundwater level. For the purpose of the study, a multilayer perceptron Artificial Neural Network (ANNs) was developed. Recently, alternative algorithms have been used for training ANNs to overcome the disadvantages of the Back Propagation (BP) algorithm that is easily stuck in local minima and slow training convergence. In this regard, the Levenberg–Marquardt algorithm as the classical method and the Particle Swarm Optimization (PSO) as the evolutionary algorithm are adopted for training the feed forward ANNs and improving their performance. The obtained results from LM-NN were then compared with those from ANN-PSO model and observed information. Comparison of the results projected by the ANN-PSO and the observed mean groundwater levels using 58 piezometric wells with monthly time steps over a 20-year period showed that the ANN-PSO model is superior to LM in predicting groundwater level. As an illustration, for models run using nine hidden neurons for Nekouabad right zones the root mean square error (RMSE) of the testing dataset for ANN-PSO was the lowest (1.50) compared to those for LM-NN (1.76). Accordingly, the ANN-PSO models are able to be used as a reliable tool for evaluating different scenarios of the water resources management in the study aquifer. Finally, three management scenarios under two climate change scenarios, A2 and B1 (obtained from GCMs), were defined and the trained ANN-PSO was subsequently used to project the effects of each scenario on the groundwater level in the plain.

Groundwater flow modelling is necessary for sustainable management of groundwater resources. Numerical and empirical models can be effectively used for modelling of groundwater flow. The specific boundary conditions, hydrogeological variables and complex aquifer structures are the pre requisite for numerical models but empirical models entirely depends on the data available for input and output parameters. This paper aims to compare the effectiveness of the numerical model using MODFLOW and empirical model Radial Basis Function Neural Network (RBFNN) developed for forecasting the groundwater levels of Athiyannoor Block Panchayath of Trivandrum district, which is categorized as semi critical zone due to the rapid decline in groundwater level. The groundwater flow model was developed with weekly groundwater level data during January 2014 to December 2014. Model was calibrated using trial and error method and groundwater levels at 10 observation wells were simulated. Using the simulated model, groundwater levels were predicted and validated from January 2015 to March 2015. The inputs to the RBFNN model includes weekly groundwater recharge, evapotranspiration, pumping rate in the pumping wells and groundwater levels in these wells at the previous time step. The trained RBFNN model is then validated. The predicted groundwater levels by numerical model and RBFNN models were compared with the observed groundwater levels during the validation period. The performance characteristics of both models indicate that RBFNN model is better than numerical model using MODFLOW for weekly groundwater level forecasting.

Journal of Hydroinformatics, 2008

A module that uses neural networks was developed for forecasting the groundwater changes in an aquifer. A modified standard Feedforward Neural Network (FNN), trained with the Levenberg–Marquardt (LM) algorithm with five input variables (precipitation, temperature, runoff, groundwater level and specific yield) with a deterministic component, is used. The deterministic component links precipitation with the seasonal recharge of the aquifer and projects the seasonal average precipitations. A new algorithm is applied to forecast the groundwater level changes in Messara Valley, Crete, Greece, where groundwater level has been steadily decreasing due to overexploitation during the last 20 years. Results from the new algorithm show that the introduction of specific yield improved the groundwater level forecasting marginally but the linearly projected precipitation component drastically increased the window of forecasting up to 30 months, equivalent to five biannual time-steps.

Resources Environment and Information Engineering, 2020

Citation: Javadinejad S, Dara R and Jafary F. How groundwater level can predict under the effect of climate change by using artificial neural networks of NARX. Resour Environ Inf Eng, 2020, 2(1): 90-99. Abstract: The phenomenon of climate change in recent years has led to significant changes in climatic elements and as a result the status of surface and groundwater resources, especially in arid and semi-arid regions, this issue has sometimes caused a significant decline in groundwater resources. In this paper, the effects of climate change on the status of groundwater resources in Marvdasht plain have been investigated. Water supply of different parts of this region is highly dependent on groundwater resources and therefore the study of groundwater changes in future periods is important in the development of this plain and the management of its water resources. In order to evaluate the effects of climate change, the output of atmospheric circulation models (GCM) has been used. Then, in order to adapt the output scale of these models to the scale required by local studies of climate change, precipitation and temperature data have been downscaled by LARS-WG model. Downscaled information was used to determine the amount of feed and drainage of the aquifer in future periods. To investigate changes in groundwater levels at different stages, a neural network dynamic model has been developed in MATLAB software environment. It is also possible to study and compare other points using other scenarios and mathematical modelling. The results of the study, assuming the current state of development in the region, indicate a downward trend in the volume of the aquifer due to climate change and its effects on resources and uses of the study area. The results also introduce Scenario A2 as the most critical scenario related to climate change, which also shows the largest aquifer decline in neural network modelling.

Journal of Water and Climate Change , 2022

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.

Journal of Water and Climate Change

The purpose of this study is the projection of climate change's impact on the Groundwater Level (GWL) fluctuations in the Mashhad aquifer during the future period (2022–2064). In the first step, the climatic variables using ACCESS-CM2 model under the Shared Socio-economic Pathways (SSPs) 5–8.5 scenario were extracted. In the second step, different machine learning algorithms, including Multilayer Perceptron Neural Network (MLP), Adaptive Neuro-fuzzy Inference System Neutral Network (ANFIS), Radial Basis Function Neural Network (RBF), and Support Vector Machine (SVM) were employed for the GWL fluctuations time series prediction under climate change in the future. Our results point out that temperatures and evaporation will increase in the autumn season, and precipitation will decrease by 26%. The amount of evaporation will increase in the winter due to an increase in temperature and a decrease in precipitation. The results showed that the RBFNN model had an excellent performance ...

Computers & Geosciences, 2013

The knowledge of groundwater table fluctuations is important in agricultural lands as well as in the studies related to groundwater utilization and management levels. This paper investigates the abilities of Gene Expression Programming (GEP), Adaptive Neuro-Fuzzy Inference System (ANFIS), Artificial Neural Networks (ANN) and Support Vector Machine (SVM) techniques for groundwater level forecasting in following day up to 7-day prediction intervals. Several input combinations comprising water table level, rainfall and evapotranspiration values from Hongcheon Well station (South Korea), covering a period of eight years (2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008) were used to develop and test the applied models. The data from the first six years were used for developing (training) the applied models and the last two years data were reserved for testing. A comparison was also made between the forecasts provided by these models and the Auto-Regressive Moving Average (ARMA) technique. Based on the comparisons, it was found that the GEP models could be employed successfully in forecasting water table level fluctuations up to 7 days beyond data records.

Journal of Groundwater Science and Engineering, 2021

Determining the link between rainfall and flow for a watershed is one of the most imperative problems and challenging tasks faced by hydrologists and engineers. Conceptual and Box-Jenkins hydrological models represent suitable tools for this purpose in circumstance of data Scarce and climate complexity. This research consists in a comparative study between conceptual models and Box-Jenkins model, namely, GR2M, ABCD, and the autoregressive moving average (ARIMA) which has a numerical design. The three models were applied to three catchments located in the northwest of Algeria. Basins have been selected according to the availability of long-time series of hydrological and climatic data (more than 30 years) to calibrate parsimonious models, taking into account the climatic variables and the stochastic behavior of the natural stream flow. Overall, the conceptual models perform similarly; whereas the results show that the GR2M model performed better than the ABCD in the validation stage, the stochastic model shows better results as opposed to conceptual models in the case of the Mellah Wadi which presents high permeability in its behavior. This is due to the simplicity of the model needed for data (only runoff data) and the ability of the stochastic model to produce stream flow in complex catchments. Such circumstance could be caused by different motivations. On the one hand, the diverse number of model parameters that make the ABCD the less parsimonious approach, with four parameters to be calibrated. On the other hand, the inability of the ABCD and the ARIMA model to capture and describe the groundwater processes, important for the cases study. Moreover, the validation period includes a large drought period, started in the late 1980s, which makes difficult model adaptation to different hydrological regimes.