Impacts of regional characteristics on improving the accuracy of groundwater level prediction using machine learning: The case of central eastern continental United States
TL;DR: In this paper, a gated recurrent unit (GRU) neural network was built for groundwater level simulation in 78 catchments in the study region, and principal component analysis was used to cluster a variety of catchment hydrological variables and determine the input variables for the GRU model.
About: This article is published in Journal of Hydrology: Regional Studies.The article was published on 2021-10-01 and is currently open access. It has received 15 citations till now. The article focuses on the topics: Groundwater & Hydrogeology.
Citations
More filters
TL;DR: In this article , a new perspective on drought propagation using convergent cross mapping (CCM) based on pure observations was provided, indicating that causality analysis would be more powerful than correlation analysis, especially for detecting drought propagation direction.
Abstract: The essence of propagation from meteorological to hydrological drought is the cause-effect relationship between precipitation and runoff. This study challenged the reliability of applying linear or non-linear correlation (i.e., closeness/similarity, a non-directional scalar) to study drought propagation (i.e., causality, a directional vector). Meanwhile, in the field of hydrometeorology, causality analysis is burgeoning in model simulations, but still rare in analyzing the observations. Therefore, this study aims to provide a new perspective on drought propagation (i.e., causality) using convergent cross mapping (CCM) based on pure observations. Compared with the results in previous studies, the effectiveness of applying causality analysis in drought propagation study was proven, indicating that causality analysis would be more powerful than correlation analysis, especially for detecting drought propagation direction.
10 citations
TL;DR: Wang et al. as discussed by the authors analyzed the GRACE RL05 mascon data during 2002-2017 in NCP via independent component analysis and multi-source data to improve the understanding of underlying physical or human-made processes of the regional TWS change.
4 citations
TL;DR: In this article , a committee machine intelligence system (CMIS) model has been developed to predict the groundwater level with high accuracy with an AARE value of less than 0.11%.
Abstract: Increasing the depth of mining leads to the location of the mine pit below the groundwater level. The entry of groundwater into the mining pit increases costs as well as reduces efficiency and the level of work safety. Prediction of the groundwater level is a useful tool for managing groundwater resources in the mining area. In this study, to predict the groundwater level, multilayer perceptron, cascade forward, radial basis function, and generalized regression neural network models were developed. Moreover, four optimization algorithms, including Bayesian regularization, Levenberg–Marquardt, resilient backpropagation, and scaled conjugate gradient, are used to improve the performance and prediction ability of the multilayer perception and cascade forward neural networks. More than 1377 data points including 12 spatial parameters divided into two categories of sediments and bedrock (longitude, latitude, hydraulic conductivity of sediments and bedrock, effective porosity of sediments and bedrock, the electrical resistivity of sediments and bedrock, depth of sediments, surface level, bedrock level, and fault), and besides, 6 temporal parameters are used (day, month, year, drainage, evaporation, and rainfall). Also, to determine the best models and combine them, 165 extra validation data points are used. After identifying the best models from the three candidate models with a lower average absolute relative error (AARE) value, the committee machine intelligence system (CMIS) model has been developed. The proposed CMIS model predicts groundwater level data with high accuracy with an AARE value of less than 0.11%. Sensitivity analysis indicates that the electrical resistivity of sediments had the highest effect on the groundwater level. Outliers’ estimation applying the Leverage approach suggested that only 2% of the data points could be doubtful. Eventually, the results of modeling and estimating groundwater level fluctuations with low error indicate the high accuracy of machine learning methods that can be a good alternative to numerical modeling methods such as MODFLOW.
3 citations
TL;DR: Wang et al. as discussed by the authors developed 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.
Abstract: • A novel hybrid model for simulating groundwater level was developed. • The hybrid model integrated water balance equations with deep learning algorithm. • The proposed model presented the superiority and powerful simulation ability. • The automatic parameterizing ability enhanced the model for cross-region simulation. 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.
3 citations
TL;DR: In this article , a graph neural network (GNN) is used to forecast groundwater dynamics where it can represent each well as a node in the graph and extract the spatial information is extracted from an interconnected network using graph convolution layers with a self-adaptive adjacency matrix.
Abstract: 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.
2 citations
References
More filters
TL;DR: A novel, efficient, gradient based method called long short-term memory (LSTM) is introduced, which can learn to bridge minimal time lags in excess of 1000 discrete-time steps by enforcing constant error flow through constant error carousels within special units.
Abstract: Learning to store information over extended time intervals by recurrent backpropagation takes a very long time, mostly because of insufficient, decaying error backflow. We briefly review Hochreiter's (1991) analysis of this problem, then address it by introducing a novel, efficient, gradient based method called long short-term memory (LSTM). Truncating the gradient where this does not do harm, LSTM can learn to bridge minimal time lags in excess of 1000 discrete-time steps by enforcing constant error flow through constant error carousels within special units. Multiplicative gate units learn to open and close access to the constant error flow. LSTM is local in space and time; its computational complexity per time step and weight is O. 1. Our experiments with artificial data involve local, distributed, real-valued, and noisy pattern representations. In comparisons with real-time recurrent learning, back propagation through time, recurrent cascade correlation, Elman nets, and neural sequence chunking, LSTM leads to many more successful runs, and learns much faster. LSTM also solves complex, artificial long-time-lag tasks that have never been solved by previous recurrent network algorithms.
72,897 citations
TL;DR: This paper is concerned with the construction of planes of closest fit to systems of points in space and the relationships between these planes and the planes themselves.
Abstract: (1901). LIII. On lines and planes of closest fit to systems of points in space. The London, Edinburgh, and Dublin Philosophical Magazine and Journal of Science: Vol. 2, No. 11, pp. 559-572.
10,656 citations
9,050 citations
03 Sep 2014
TL;DR: In this paper, a gated recursive convolutional neural network (GRNN) was proposed to learn a grammatical structure of a sentence automatically, which performed well on short sentences without unknown words, but its performance degrades rapidly as the length of the sentence and the number of unknown words increase.
Abstract: Neural machine translation is a relatively new approach to statistical machine translation based purely on neural networks. The neural machine translation models often consist of an encoder and a decoder. The encoder extracts a fixed-length representation from a variable-length input sentence, and the decoder generates a correct translation from this representation. In this paper, we focus on analyzing the properties of the neural machine translation using two models; RNN Encoder‐Decoder and a newly proposed gated recursive convolutional neural network. We show that the neural machine translation performs relatively well on short sentences without unknown words, but its performance degrades rapidly as the length of the sentence and the number of unknown words increase. Furthermore, we find that the proposed gated recursive convolutional network learns a grammatical structure of a sentence automatically.
4,702 citations
TL;DR: This work analyzes two classes of controls consisting of patchy nucleotide sequences generated by different algorithms--one without and one with long-range power-law correlations, finding that both types of sequences are quantitatively distinguishable by an alternative fluctuation analysis method.
Abstract: Long-range power-law correlations have been reported recently for DNA sequences containing noncoding regions We address the question of whether such correlations may be a trivial consequence of the known mosaic structure ("patchiness") of DNA We analyze two classes of controls consisting of patchy nucleotide sequences generated by different algorithms--one without and one with long-range power-law correlations Although both types of sequences are highly heterogenous, they are quantitatively distinguishable by an alternative fluctuation analysis method that differentiates local patchiness from long-range correlations Application of this analysis to selected DNA sequences demonstrates that patchiness is not sufficient to account for long-range correlation properties
4,365 citations