Groundwater Level Monitoring Network Research Articles

Groundwater level monitoring network design with machine learning methods

This research introduces a method combining groundwater models and machine learning (ML) algorithms to locate observation wells and design optimal Groundwater Level Monitoring Networks (GLMNs). Groundwater models and stochastic simulations are used to extract required hydrogeological datasets for ML algorithms. In addition to data generation, the stochastic simulations minimize the uncertainties in the aquifer characterization, leading to a precise design of GLMNs. In this research, K-means clustering and relevance vector machine (RVM) are the ML algorithms employed to determine the optimal configuration of observation wells in terms of number and location in a monitoring network. This study proposes three GLMNs (K-mean, RVM, modified RVM), compares them with the existing observation wells, and investigates their effects on the accuracy of groundwater modeling and running time. The groundwater model with a K-mean network runs faster than other configurations, while the model with a modified RVM network shows a significant decrease in errors.

The Environment Agency Chalk groundwater level monitoring network in England

Abstract This paper reviews the Environment Agency chalk groundwater level monitoring network. The network has evolved over many years to enable management of the resource and to assess the impact of abstractions on the environment. The paper considers the strengths and weaknesses of the network, the use and accessibility of the data and how the network supports, and works with, the Environment Agency's regional groundwater models. It concludes with the suggestion that the network is suffering from a degree of lack of maintenance and that there is a disparity between the ambitions of the modelling programme with its Modflow6-driven shift to multi-layer conceptualization and a largely open-hole, single-layer monitoring installation.

Design of preliminary groundwater monitoring networks for the coastal Tra Vinh province in Mekong Delta, Vietnam

Study RegionThe study region is the coastal Tra Vinh province located in Mekong Delta, Vietnam Study FocusThe objective of the study is to design preliminary monitoring networks for groundwater level and salinity monitoring in the three heavily abstracted aquifers in the Tra Vinh province. Since data from the existing monitoring were insufficient to explore variogram analysis and subsequent Kriging interpolation for the network design, groundwater flow and saltwater transport models were constructed to create groundwater level and salinity distributions. An iterative procedure was adopted to select monitoring well locations based on the contour maps of groundwater levels and total dissolved solids (TDS) constructed from the groundwater models. The standard deviation of Kriging interpolation error was used as the criteria to assess the quality of the designed monitoring networks. Priorities for the locating groundwater level monitoring wells were targeted in areas with cones of depression and boundaries, while priorities for salinity monitoring were given to coastal zone and inland fresh/saltwater interfaces. New Hydrological Insights for the RegionGroundwater overexploitation has resulted in storage depletion and saltwater intrusion in the Tra Vinh province. Simulations from the numerical models determined the cones of the depression and predicted intrusion from inland saltwater trapped in clays from past marine transgressions and from seawater in the coastal zone. A limited number of monitoring wells is operated and is insufficient to delineate the cone of depression and fresh/salt water interfaces. This study combined the numerical model simulation results and Kriging interpolation method and designed preliminary groundwater level and salinity monitoring networks for the Tra Vinh province. Once the designed networks are implemented, observed data from the networks will provide valuable information for sustainable groundwater resources management in Tra Vinh province. Data availabilityAll data generated or analyzed during this study are available upon request from the corresponding author.

Groundwater level monitoring of Danish chalk and limestone aquifers to survey trends in quantitative status and groundwater resources

Abstract Chalk and limestone aquifers contribute one-third of the drinking water supply in Denmark, and one-sixth of that national groundwater resource is assessed as having a ‘Poor’ status in terms of the quantitative Water Framework Directive due to intensive abstraction. This paper describes the national groundwater level monitoring network with regard to the following three applications: (1) when used for the annual surveying and reporting of groundwater resources and impacts from climate and groundwater abstraction; (2) as part of real-time monitoring and modelling for daily and seasonal forecasting; and (3) for tracking long-term climate change impacts on groundwater levels. Groundwater level monitoring provides a particularly important indicator of abstraction pressure and sustainable balance compared with recharge. Many larger chalk and limestone groundwater bodies in Denmark are only monitored by local water companies and not represented in the national groundwater level network. This raises the concern that current national groundwater level monitoring does not fully support integrated modelling and assessment purposes for chalk and limestone groundwater bodies. This also implies that, for tracking long-term climate change and anthropogenic impacts on groundwater levels, the national groundwater monitoring network especially lacks long-term records with complete 30-year time-series for many intensively exploited large chalk and limestone aquifers.

Spatial Variability in Land Subsidence and Its Relation to Groundwater Withdrawals in the Choshui Delta

In this article, the characterization of land subsidence with the spatial variability of soil formation and groundwater withdrawals in Choshui delta, Taiwan, is presented. Levelling surveys, borehole logging, multi-layer compaction monitoring network, multi-layer groundwater level monitoring network, and the electricity consumption of wells in the study area are adopted. Various factors, including the percentage of coarse-grained soil, percentage of fine-grained soil, length of average maximum drainage path, percentage of agricultural land use, electricity consumption of wells, and accumulated depth of land subsidence, are applied. Thematic maps based on these factors are established using geographic information system spatial analysis. A principal component analysis (PCA) is then employed to obtain the dominant factors for land subsidence. The results indicate that the largest subsidence rate is observed in the region that has both a high electricity consumption of wells and a large percentage of fine-grained soil. The PCA results reveal that the electricity consumption of wells is highly correlated with the accumulated depth of land subsidence. The first principal component is the acquired factor causing land subsidence, such as the excessive withdrawal of groundwater. The second principal component is the congenital factor causing land subsidence, which corresponds to fine sand, silty and clayey soils.

Optimizing groundwater level monitoring networks with hydrogeological complexity and grid-based mapping methods

Optimizing groundwater level monitoring networks with hydrogeological complexity and grid-based mapping methods

Evaluation of two linear kriging methods for piezometric levels interpolation and a framework for upgrading groundwater level monitoring network in Ghiss-Nekor plain, north-eastern Morocco

Evaluation of two linear kriging methods for piezometric levels interpolation and a framework for upgrading groundwater level monitoring network in Ghiss-Nekor plain, north-eastern Morocco

A decision support tool for optimising groundwater-level monitoring networks using an adaptive genetic algorithm

Abstract Mapping of the spatial variability of sparse groundwater-level measurements is usually achieved by means of geostatistical methods. This work tackles the problem of deficient sampling of an aquifer, by employing an innovative integer adaptive genetic algorithm (iaGA) coupled with geostatistical modelling by means of ordinary kriging, to optimise the monitoring network. Fitness functions based on three different errors are used for removing a constant number of boreholes from the monitoring network. The developed methodology has been applied to the Mires basin in Crete, Greece. The methodological improvement proposed concerns the adaptive method for the GA, which affects the crossover–mutation fractions depending on the stall parameter, aiming at higher accuracy and faster convergence of the GA. The initial dataset consists of 70 monitoring boreholes and the applied methodology shows that as many as 40 boreholes can be removed, while still retaining an accurate mapping of groundwater levels. The proposed scenario for optimising the monitoring network consists of removing 30 boreholes, in which case the estimated uncertainty is considerably smaller. A sensitivity analysis is conducted to compare the performance of the standard GA with the proposed iaGA. The integrated methodology presented is easily replicable for other areas for efficient monitoring networks design.

Assessment of Groundwater-Level Monitoring Network in Irrigated Regions with a Complex Aquifer System Using Information Theory

AbstractThe present study focuses on the assessment of groundwater level monitoring networks (GWLMN) by examining the statistical relationship between groundwater level (GWL) uncertainty and sustai...

Building a Low‐Cost, Internet‐of‐Things, Real‐Time Groundwater Level Monitoring Network

AbstractA community‐based, real‐time, groundwater level monitoring network consisting of 11 sites was built in Nova Scotia, Canada, using privately owned domestic wells and low‐cost, custom‐made water level meters. The real‐time meters use an ultrasonic sensor to measure water levels and an Internet‐of‐Things device to transmit the data to the Internet by WiFi or cellular connection. The water level data are plotted in real‐time on a time‐series graph and are available immediately for online viewing and downloading. Based on observations at three sites, the real‐time water level meter data compare well to pressure transducer measurements, with mean absolute errors of less than 0.02 m. The meters are simple to build, and components are readily available from online suppliers at low cost.

Modelling Groundwater Hydraulics to Design a Groundwater Level Monitoring Network for Sustainable Management of Fresh Groundwater Lens in Lower Indus Basin, Pakistan

The over-extraction of groundwater from thin fresh groundwater lenses is a threat to the livelihood of farmers in the Lower Indus Basin (LIB). It is essential to monitor and regulate this pumping to sustain fresh groundwater lenses. In this study, we applied a modelling approach in combination with geostatistical analysis to identify the critical locations to monitor the groundwater levels for sustaining fresh groundwater in the LIB. Our approach included four steps: (i) simulating temporal heads using a calibrated hydrogeological model; (ii) sampling monitoring locations using a hexagonal pattern of sampling; (iii) applying principal component analysis (PCA) of the temporal head observations, and selecting high scoring locations from the PCA; and (iv) minimizing the observation points to represent the water level contours. The calibrated model was able to replicate the hydro-dynamic behavior of the study area, with a root mean square of 0.95 and an absolute residual mean of 0.74 m. The hexagonal pattern of spatial sampling resulted in a 195 point network, but PCA reduced this network to 135 points and contour classification reduced it even further to 59 points. The 195, 135, and 59 point networks represented the water levels with average standard errors of 0.098, 0.318, and 0.610 m, respectively. Long-term simulations with increased pumping showed that the water levels would best be assessed by 195 monitoring points, although 135 and 59 points would represent the depleting area but would not capture the water logging area.

Design of the groundwater level monitoring network using principal component analysis (PCA) technique.

Using an appropriate monitoring network is considered as an efficient option to manage the groundwater resources and reduce drilling of costly sampling wells. Principal component analysis (PCA) is one of the data reduction techniques used to extract essential components. The used techniques are based on the identification of those describing the variance of the system. In this paper, the PCA technique has been employed in order to identify the effective wells and remove the less important ones. For this purpose, 160 wells were constructed in the Salman Farsi Agro-Industry, located in Khuzestan province of Iran. The data are measured twice a month for 12 months. In this technique, variation factors called principal components are identified through considering the data structures. Using the PCA, the relative importance of each well has been calculated for the groundwater depth estimation. In the present study, the acceptable threshold has been taken to be 0.8 and therefore the number of wells in determining groundwater depth was reduced to 33 ones. Identifying the essential wells, the important points for sampling are identified and groundwater depth monitoring is performed only in these wells. This will save time and cost of groundwater level monitoring within the study area.

A Review of the Historical Background, Needs, Design approaches and Future Challenges in Groundwater Level Monitoring Networks

A Review of the Historical Background, Needs, Design approaches and Future Challenges in Groundwater Level Monitoring Networks

Analysis of the long-term effects of groundwater extraction on the water balance in part of the Urucuia Aquifer System in Bahia - Brazil

In agricultural regions where there is insufficient rainwater for cultivation, understanding the dynamics of surface water and groundwater is critical to assess the impact of increased well pumping on the water balance. The western region of the state of Bahia-Brazil, the largest area of agribusiness in the state, has experienced progressive occupation since the 1980s, resulting in pressure on water resources - mainly after the introduction of irrigation - and conflicts among water users. This study analyzed the effects of groundwater extraction by wells in a portion of the Urucuia Aquifer System. The methodology used was the simulation of groundwater flow in steady-state for three scenarios: i) without withdrawal of water by pumping wells; (ii) with current withdrawal rates; and (iii) with 60% additional extraction. After defining well production rates from field surveys (4.6 m3 s-1) and modeling of the aquifer base geometry (maximum thickness of 535 m), the steady-state models with and without pumping show that, in the long term, groundwater storage decreases by 2 km3 (0.8 m on average in the aquifer) without interference in the regional flow direction. The mass balance shows that the base flow of the main rivers is reduced by approximately 6% after current groundwater extraction and an additional 2.5% after additional extraction. The results point to a greater impact on surface water caused by a reduction in groundwater storage and related river base flow. This indicates the importance of maintaining and expanding the groundwater-level monitoring network.

On the Optimal Spatial Design for Groundwater Level Monitoring Networks

AbstractEffective groundwater monitoring networks are important, as systematic data collected at observation wells provide a crucial understanding of the dynamics of hydrogeological systems as well as the basis for many other applications. This study investigates the influence of six groundwater level monitoring network (GLMN) sampling designs (random, grid, spatial coverage, and geostatistical) with varying densities on the accuracy of spatially interpolated groundwater surfaces. To obtain spatially continuous prediction errors (in contrast to point cross‐validation errors), we used nine potentiometric groundwater surfaces from three regional MODFLOW groundwater flow models with different resolutions as a priori references. To assess the suitability of frequently‐used cross‐validation error statistics (MAE, RMSE, RMSSE, ASE, and NSE), we compared them with the actual prediction errors (APE). Additionally, we defined upper and lower thresholds for an appropriate spatial density of monitoring wells. Below the lower threshold, the observation density appears insufficient, and additional wells lead to a significant improvement of the results. Above the upper threshold, additional wells lead to only minor and inefficient improvements. According to the APE, systematic sampling lead to the best results but is often not suited for GLMN due to its nonprogressive characteristic. Geostatistical and spatial coverage sampling are considerable alternatives, which are in contrast progressive and allow evenly spaced and, in the case of spatial coverage sampling, yet reproducible coverage with accurate results. We found that the global cross‐validation error statistics are not suitable to compare the performance of different sampling designs, although they allow rough conclusions about the quality of the GLMN.

Improving the reliability of groundwater monitoring networks using combined numerical, geostatistical and neural network-based simulation models

ABSTRACTA new methodology is proposed for improving the accuracy of groundwater-level estimations and increasing the efficiency of groundwater-level monitoring networks. Three spatio-temporal (S-T) simulation models, numerical groundwater flow, artificial neural network and S-T kriging, are implemented to simulate water-table level variations. Individual models are combined using model fusion techniques and the more accurate of the individual and combined simulation models is selected for the estimation. Leave-one-out cross-validation shows that the estimation error of the best fusion model is significantly less than that of the three individual models. The selected fusion model is then considered for optimal S-T redesign of the groundwater monitoring network of the Dehgolan Plain (Iran). Using a Bayesian maximum entropy interpolation technique, soft data are included in the geostatistical analyses. Different scenarios are defined to incorporate economic considerations and different levels of precision in selecting the best monitoring network; a network of 37 wells is proposed as the best configuration. The mean variance estimation errors of all scenarios decrease significantly compared to that of the existing monitoring network. A reduction in equivalent uniform annual costs of different scenarios is achieved.

Trend prediction of regional groundwater level with GISFEFLOW model in Beijing Mihuaishun plain, China

A three dimensional finite element model (FEFLOW) has been used for regional ground water flow modeling of Mihuaishun plain located in the south of Beijing, China. The numerical groundwater flow model is developed considering recharge components (precipitation infiltration, river leakage, and irrigation return flow). GIS interface is created for each source of recharge. Hydraulic conductivities and storage coefficient have been calibrated by the steady state model using the recorded data from 2007 to 2009. The model results is useful to identify the aquifer characteristics and to analyze the groundwater dynamics. The groundwater level monitoring network will be improved by analyzing groundwater levels. The future development scenarios are proposed to predict the trend changes of groundwater levels.

Influence of available data on the geostatistical-based design of optimal spatiotemporal groundwater-level-monitoring networks

Several methodologies exist for the design of optimal groundwater monitoring networks. Those that employ geostatistical techniques are usually dependent on a variogram. In practice, different variograms for an environmental variable can be fitted depending on the number of available data, their spatial distribution, and their variability. For this reason, in most cases, it is difficult to get an adequate variogram. This report evaluates the influence of the groundwater level data available in the construction of a product-sum spatiotemporal variogram model and its consequences on a geostatistical-based methodology that uses the Kalman filter for the design of optimal spatiotemporal monitoring networks. Additionally, a sensitivity analysis is carried out to determine the variations in the optimal number of wells and monitoring frequency for the monitoring network, depending on the selected spatiotemporal variogram model. The results show the importance of getting a good quality spatiotemporal variogram when designing optimal groundwater-level monitoring networks using geostatistical-based methodologies.

Optimal design of groundwater-level monitoring networks

Abstract Groundwater monitoring plays a significant role in groundwater management. This study presents an optimization method for designing groundwater-level monitoring networks. The proposed design method was used in the Eshtehard aquifer, in central Iran. Three scenarios were considered to optimize the locations of the observation wells: (1) designing new monitoring networks, (2) redesigning existing monitoring networks, and (3) expanding existing monitoring networks. The kriging method was utilized to determine groundwater levels at non-monitoring locations for preparing the design data base. The optimization of the groundwater monitoring network had the objectives of (1) minimizing the root mean square error and (2) minimizing the number of wells. The non-dominated sorting genetic algorithm (NSGA-II) was applied to optimize the network. Inverse distance weighting interpolation was used in NSGA-II to estimate the groundwater levels while optimizing network design. Results of the study indicate that the proposed method successfully optimizes the design of groundwater monitoring networks that achieve accuracy and cost-effectiveness.

A Bayesian maximum entropy-based methodology for optimal spatiotemporal design of groundwater monitoring networks.

This paper presents a new methodology for analyzing the spatiotemporal variability of water table levels and redesigning a groundwater level monitoring network (GLMN) using the Bayesian Maximum Entropy (BME) technique and a multi-criteria decision-making approach based on ordered weighted averaging (OWA). The spatial sampling is determined using a hexagonal gridding pattern and a new method, which is proposed to assign a removal priority number to each pre-existing station. To design temporal sampling, a new approach is also applied to consider uncertainty caused by lack of information. In this approach, different time lag values are tested by regarding another source of information, which is simulation result of a numerical groundwater flow model. Furthermore, to incorporate the existing uncertainties in available monitoring data, the flexibility of the BME interpolation technique is taken into account in applying soft data and improving the accuracy of the calculations. To examine the methodology, it is applied to the Dehgolan plain in northwestern Iran. Based on the results, a configuration of 33 monitoring stations for a regular hexagonal grid of side length 3600m is proposed, in which the time lag between samples is equal to 5weeks. Since the variance estimation errors of the BME method are almost identical for redesigned and existing networks, the redesigned monitoring network is more cost-effective and efficient than the existing monitoring network with 52 stations and monthly sampling frequency.