Use of GIS,Statistics and Machine Learning for Groundwater Quality Management: Application to Nitrate Contamination

Water Resources - Groundwater NO3 contamination (GNC) threatens the drinkability of water in many countries worldwide. It could cause serious health problems and sometimes lead to death. This paper...     

A Simple Relation to Constrain Groundwater Models Using Surface Deformation

A simple relation between pore pressure change and one-dimensional surface deformation is presented. The relation is for pore pressure change in a confined aquifer that causes surface deformation. It can be applied to groundwater models of any discretization and is computationally efficient. The estimated surface deformation from model results can be compared to observed surface deformation through geodetic techniques such as Differential Interferometric Synthetic Aperture Radar. Model parameters then are constrained using the observed surface deformation. The validity of this relation is shown through constraint of model parameters for surface uplift due to pore pressure increase caused by wastewater disposal injection.     

Forecasting Water Levels in Krasnodar Krai Rivers with the Use of Machine Learning

Water Resources - The potential of construction of machine learning models was considered as applied to water level forecasting in mountain river reaches in Krasnodar Krai based on observation data...     

Permeable Asphalt: A New Tool to Reduce Road Salt Contamination of Groundwater in Urban Areas

Chloride contamination of groundwater in urban areas due to deicing is a well‐documented phenomenon in northern climates. The objective of this study was to evaluate the effects of permeable pavement on degraded urban groundwater. Although low impact development practices have been shown to improve stormwater quality, no infiltration practice has been found to prevent road salt chlorides from entering groundwater. The few studies that have investigated chlorides in permeable asphalt have involved sampling directly beneath the asphalt; no research has looked more broadly at surrounding groundwater conditions. Monitoring wells were installed upgradient and downgradient of an 860 m² permeable asphalt parking lot at the University of Connecticut (Storrs, Connecticut). Water level and specific conductance were measured continuously, and biweekly samples were analyzed for chloride. Samples were also analyzed for sodium (Na), calcium (Ca), and magnesium (Mg). Analysis of variance analysis indicated a significantly (p < 0.001) lower geometric mean Cl concentration downgradient (303.7 mg/L) as compared to upgradient (1280 mg/L). Concentrations of all alkali metals increased upgradient and downgradient during the winter months as compared to nonwinter months, indicating that cation exchange likely occurred. Despite the frequent high peaks of chloride in the winter months as well as the increases in alkali metals observed, monitoring revealed lower Cl concentrations downgradient than upgradient for the majority of the year. These results suggest that the use of permeable asphalt in impacted urban environments with high ambient chloride concentrations can be beneficial to shallow groundwater quality, although these results may not be generalizable to areas with low ambient chloride concentrations.     

Application of Machine Learning Methods in Arrival Time Picking of P Waves from Reservoir Earthquakes

Reservoir earthquake characteristics such as small magnitude and large quantity may result in low monitoring efficiency when using traditional methods. However, methods based on deep learning can discriminate the seismic phases of small earthquakes in a reservoir and ensure rapid processing of arrival time picking. The present study establishes a deep learning network model combining a convolutional neural network (CNN) and recurrent neural network (RNN). The neural network training uses the waveforms of 60 000 small earthquakes within a magnitude range of 0.8-1.2 recorded by 73 stations near the Dagangshan Reservoir in Sichuan Province as well as the data of the manually picked P-wave arrival time. The neural network automatically picks the P-wave arrival time, providing a strong constraint for small earthquake positioning. The model is shown to achieve an accuracy rate of 90.7% in picking P waves of microseisms in the reservoir area, with a recall rate reaching 92.6% and an error rate lower than 2%. The results indicate that the relevant network structure has high accuracy for picking the P-wave arrival times of small earthquakes, thus providing new technical measures for subsequent microseismic monitoring in the reservoir area.     

Methods to Estimate Source Zone Depletion of Fuel Releases by Groundwater Flow

We compare two methods for estimating the natural source zone depletion (NSZD) rate at fuel release sites that occurs by groundwater flow through the source zone due to dissolution and transport of biodegradation products. Dissolution is addressed identically in both methods. The “mass budget method”, previously proposed and applied by others, estimates the petroleum hydrocarbon biodegradation rate based on dissolved electron acceptor delivery and dissolved biodegradation product removal by groundwater flow. The mass budget method relies on assumed stoichiometry for the degradation reactions and differences in concentrations of dissolved species (oxygen, nitrate, sulfate, reduced iron, reduced manganese, nonvolatile dissolved organic carbon, methane) at monitoring locations upgradient and downgradient of the source zone. We illustrate a refinement to account for degradation reactions associated with loss of reduced iron from solution. The “carbon budget method,” a simplification of approaches applied by others, addresses carbon‐containing species in solution or lost from solution (precipitated) and does not require assumptions about stoichiometry or information about electron acceptors. We apply both methods to a fuel release site with unusually detailed monitoring data and discuss applicability to more typical and less thoroughly monitored sites. The methods, as would typically be applied, yield similar results but have different constraints and uncertainties. Overall, we conclude that the carbon budget method has greater practical utility as it is simpler, requires fewer assumptions, accounts for most iron‐reducing reactions, and does not include CO₂ that escapes from the saturated to the unsaturated zone.     

Use of an Artificial Sweetener to Identify Sources of Groundwater Nitrate Contamination

The artificial sweetener acesulfame (ACE) is a potentially useful tracer of waste water contamination in groundwater. In this study, ACE concentrations were measured in waste water and impacted groundwater at 12 septic system sites in Ontario, Canada. All samples of septic tank effluent (n = 37) had ACE >6 µg/L, all samples of groundwater from the proximal plume zones (n = 93) had ACE >1 µg/L and, almost all samples from the distal plume zones had ACE >2 µg/L. Mean mass ratios of total inorganic nitrogen/ACE at the 12 sites ranged from 680 to 3500 for the tank and proximal plume samples. At five sites, decreasing ratio values in the distal zones indicated nitrogen attenuation. These ratios were applied to three aquifers in Canada that are nitrate‐stressed and an urban stream where septic systems are present nearby to estimate the amount of waste water nitrate contamination. At the three aquifer locations that are agricultural, low ACE values (<0.02‐0.15 µg/L) indicated that waste water contributed <15% of the nitrate in most samples. In groundwater discharging to the urban stream, much higher ACE values (0.2‐11 µg/L) indicated that waste water was the likely source of >50% of the nitrate in most samples. This study confirms that ACE is a powerful tracer and demonstrates its use as a diagnostic tool for establishing whether waste water is a significant contributor to groundwater contamination or not.     

从误差观点综述分析地震定位方法

阐述了现有几种地震定位方法的原理,然后从误差的观点对它们进行了定性的比较,认为本文提到的几种定位方法都是从某一方面降低误差来源的影响,它们之间不存在优劣之分,并可以结合各种方法的优点获取高精度的定位结果。     

Interface Treatment of Petroleum Hydrocarbon‐impacted Lower Permeability Layers by Activated Sodium Persulfate to Reduce Emissions to Groundwater

Nonaqueous phase liquid (NAPL)‐impacted lower permeability layers in heterogeneous media are difficult to fully remediate and can act as persistent sources of groundwater contamination through diffusive emissions to transmissive aquifer zones. This work investigated the benefits of partial remediation involving treatment focused near the high‐low permeability interface, with the performance metric being emissions reduction. A sequential base‐activated persulfate (S₂O₈ ^2?) delivery treatment strategy was studied in this work, involving 13–14 d deliveries of 10% w/w sodium persulfate (Na₂S₂O₈) and 14–28 d deliveries of 19 g/L sodium hydroxide (NaOH). Treatment and control experiments were conducted in 1.2‐m wide × 1.2‐m tall × 5‐cm thick physical model tanks containing two soil layers differing in hydraulic conductivity by three orders of magnitude. The top 10 cm of the lower permeability layers contained 7400–7800 mg‐NAPL/kg‐soil; the NAPL was comprised of benzene, toluene, ethylbenzene, p‐xylene, o‐xylene, n‐propylbenzene, and 1,3,5‐trimethylbenzene (TMB) mixed in octane. Approximately 0.1 g‐Na₂S₂O₈ was delivered per cm²‐interface area over 13–14 d. The S₂O₈ ^2? and SO₄ ^2? concentration profiles suggest higher oxidant reaction rates when NaOH is delivered prior to, rather than after Na₂S₂O₈. After 264 d and two treatments, hydrocarbon emissions from the NAPL source were reduced by 60% to 73% compared to a no‐treatment control tank. The incremental benefit of the second treatment was only about 10% of the effect of the first treatment.     

Use of Groundwater Levels with the PULSE Analytical Model

The PULSE analytical model, which calculates daily groundwater discharge on the basis of user‐specified recharge, was originally developed for calibration using streamflow data. This article describes a model application in which groundwater level data constitute the primary control on model input. As a test case, data were analyzed from a small basin in central Pennsylvania in which extensive groundwater level data are available. The timing and intensity of daily water‐level rises are used to ascertain temporal distribution of recharge, and the simulated groundwater discharge hydrograph has shape features that are similar to the streamflow hydrograph. This article does not include details about calibration, but some steps are illustrated and general procedures are described for calibration in specific hydrologic studies. The PULSE model can be used to assess results of fully automated base flow methods and can be used to define groundwater recharge and discharge at a relatively small time scale.     

Mapping Groundwater Potential Through an Ensemble of Big Data Methods

Groundwater resources are crucial to safe drinking supplies in sub-Saharan Africa, and will be increasingly relied upon in a context of climate change. The need to better understand groundwater calls for innovative approaches to make the best out of the existing information. A methodology to map groundwater potential based on an ensemble of machine learning classifiers is presented. A large borehole database (n = 1848) was integrated into a Geographic Information Systems (GIS) environment and used to train, validate and test 12 machine learning algorithms. Each classifier predicts a binary target (positive or negative borehole) based on the minimum flow rate required for communal domestic supplies. Classification is based on a number of explanatory variables, including landforms, lineaments, soil, vegetation, geology and slope, among others. Correlations between the target and explanatory variables were then generalized to develop groundwater potential maps. Most algorithms attained success rates between 80% and 96% in terms of test score, which suggests that the outcomes provide an accurate picture of field conditions. Statistical learners were observed to perform better than most other algorithms, excepting random forests and support vector machines. Furthermore, it is concluded that the ensemble approach provides added value by incorporating a measure of uncertainty to the results. This technique may be used to rapidly map groundwater potential for rural supply or humanitarian emergencies in areas where there is sufficient historical data but where comprehensive field work is unfeasible.     

Evaluating Groundwater Interbasin Flow Using Multiple Models and Multiple Types of Data

The groundwater interbasin flow, Q_y, from the north of Yucca Flat into Yucca Flat simulated using the Death Valley Regional Flow System (DVRFS) model greatly exceeds assessments obtained using other approaches. This study aimed to understand the reasons for the overestimation and to examine whether the Q_y estimate can be reduced. The two problems were tackled from the angle of model uncertainty by considering six models revised from the DVRFS model with different recharge components and hydrogeological frameworks. The two problems were also tackled from the angle of parametric uncertainty for each model by first conducting Morris sensitivity analysis to identify important parameters and then conducting Monte Carlo simulations for the important parameters. The uncertainty analysis is general and suitable for tackling similar problems; the Morris sensitivity analysis has been utilized to date in only a limited number of regional groundwater modeling. The simulated Q_y values were evaluated by using three kinds of calibration data (i.e., hydraulic head observations, discharge estimates, and constant‐head boundary flow estimates). The evaluation results indicate that, within the current DVRFS modeling framework, the Q_y estimate can only be reduced to about half of the original estimate without severely deteriorating the goodness‐of‐fit to the calibration data. The evaluation results also indicate that it is necessary to develop a new hydrogeological framework to produce new flow patterns in the DVRFS model. The issues of hydrogeology and boundary flow are being addressed in a new version of the DVRFS model planned for release by the U.S. Geological Survey.     

Multimodel Validity Assessment of Groundwater Flow Simulation Models Using Area Metric Approach

We demonstrate the application of the Area Metric developed by Ferson et al. (2008) for multimodel validity assessment. The Area Metric quantified the degree of models' replicative validity: the degree of agreement between the observed data and the corresponding simulated outputs represented as their empirical cumulative distribution functions (ECDFs). This approach was used to rank multiple representations of a case study groundwater flow model of a landfill by their Area Metric scores. A multimodel approach allows to account for uncertainties that may either be epistemic (from lack of knowledge) or aleatory (from variability inherent in the system). The Area Metric approach enables explicit incorporation of model uncertainties, epistemic as well as aleatory, into validation assessment. The proposed approach informs understanding of the collected data and that of the model domain. It avoids model overfitting to a particular system state, and in fact is a blind assessment of the models' validity: models are not adjusted, or updated, to achieve a better numerical fit. This approach assesses the degree of models' validity, in place of the typical binary model validation/invalidation process. Collectively, this increases confidence in the model's representativeness that, in turn, reduces risk to model users.     

Evaluation of Hydraulic Conductivity Estimates from Various Approaches with Groundwater Flow Models

Significant efforts have been expended for improved characterization of hydraulic conductivity (K) and specific storage (S_s) to better understand groundwater flow and contaminant transport processes. Conventional methods including grain size analyses (GSA), permeameter, slug, and pumping tests have been utilized extensively, while Direct Push-based Hydraulic Profiling Tool (HPT) surveys have been developed to obtain high-resolution K estimates. Moreover, inverse modeling approaches based on geology-based zonations, and highly parameterized Hydraulic Tomography (HT) have also been advanced to map spatial variations of K and S_s between and beyond boreholes. While different methods are available, it is unclear which one yields K estimates that are most useful for high resolution predictions of groundwater flow. Therefore, the main objective of this study is to evaluate various K estimates at a highly heterogeneous field site obtained with three categories of characterization techniques including: (1) conventional methods (GSA, permeameter, and slug tests); (2) HPT surveys; and (3) inverse modeling based on geology-based zonations and highly parameterized approaches. The performance of each approach is first qualitatively analyzed by comparing K estimates to site geology. Then, steady-state and transient groundwater flow models are employed to quantitatively assess various K estimates by simulating pumping tests not used for parameter estimation. Results reveal that inverse modeling approaches yield the best drawdown predictions under both steady and transient conditions. In contrast, conventional methods and HPT surveys yield biased predictions. Based on our research, it appears that inverse modeling and data fusion are necessary steps in predicting accurate groundwater flow behavior.     

Use of Groundwater Level Fluctuations near an Operating Water Supply Well to Estimate Aquifer Transmissivity

We developed a method to estimate aquifer transmissivity from the hydraulic-head data associated with the normal cyclic operation of a water supply well thus avoiding the need for interrupting the water supply associated with a traditional aquifer test. The method is based on an analytical solution that relates the aquifer's transmissivity to the standard deviation of the hydraulic-head fluctuations in one or more observation wells that are due to the periodic pumping of the production well. We analyzed the resulting analytical solution and demonstrated that when the observation wells are located near the pumping well, the solution has a simple, Dupuit like form. Numerical analysis demonstrates that the analytical solution can also be used for a quasi-periodic pumping of the supply well. Simulation of cyclic pumping in a statistically heterogeneous medium confirms that the method is suitable for analyzing the transmissivity of weakly or moderately heterogeneous aquifers. If only one observation well is available, and the shift in the phase of hydraulic-head oscillations between the pumping well and the observation well is not identifiable. Prior knowledge of aquifer's hydraulic diffusivity is required to obtain the value of the aquifer transmissivity.     

机器学习在地震预测中的应用进展

机器学习(Machine Learning, ML),特别是深度学习(Deep Learning, DL),在最近几年发展迅速,在数据挖掘、计算机视觉、自然语言处理、数据特征提取和预测等方面的应用中取得了令人振奋的进展。地震预测是复杂、涉及面广、不成熟而且充满争议的科学问题;其发展受到尚不清楚的地震机理和孕震结构、不完备的观测数据与真伪不清的地震现象等方面的限制。但是,机器学习有可能改善复杂地震数据的挖掘和发现,推动地震预测科学的发展。本文回顾了机器学习在地震预测的应用,包括强震、强余震和岩石破裂失稳等方面的预测,并展望了机器学习在地震预测方面的研究趋势。     

Application of Groundwater Vulnerability Overlay and Index Methods to the Jijel Plain Area (Algeria)

Today, scientists are deeply concerned by the vulnerability of groundwater reservoirs to pollution. Relatively simple overlay and index methods can be used to produce groundwater vulnerability maps in geographic information system. In addition, this study deals with contamination from nonpoint sources. In this study, two such models, DRASTIC and GOD, were applied in the Jijel Plain area of northeast Algeria and compared with measured groundwater nitrate concentrations. This showed that results from DRASTIC were better than GOD, 69% correlation with nitrate compared to 56%. DRASTIC was better able to identify vulnerable zones along the river valleys. The DRASTIC model was then modified using the nitrate concentrations to optimize the rating score given within each parameter range and sensitivity analysis to change the weighting given for each parameter. These combined changes gave a final Pearson's correlation of 83% with nitrate. This showed that recharge, aquifer type, and topography were the key factors in controlling vulnerability to nitrate pollution.