Ensemble Boosting and Bagging Based Machine Learning Models for Groundwater Potential Prediction

Water Resources Management - Due to the rapidly increasing demand for groundwater, as one of the principal freshwater resources, there is an urge to advance novel prediction systems to more...     

Quadratic Discriminant Analysis Based Ensemble Machine Learning Models for Groundwater Potential Modeling and Mapping

In this study, the AdaBoost, MultiBoost and RealAdaBoost methods were combined with the Quadratic Discriminant Analysis method to develop three new GIS-based Machine Learning ensemble models, i.e., ABQDA, MBQDA, and RABQDA for groundwater potential mapping in the Dak Nong Province, Vietnam. In total, 227 groundwater wells and 12 conditioning factors (infiltration, rainfall, river density, topographic wetness index, sediment transport index, stream power index, elevation, aspect, curvature, slope, soil, and land use) were used for this study. Performance of the models was evaluated using the Area Under the Receiver Operating Characteristics Curve AUC (AUC) and several other performance metrics. The results showed that the ABQDA model that achieved AUC?=?0.741 was superior to the other models in producing an accurate map of groundwater potential for the Dak Nong Province. The models and potential maps produced here can help policymakers and water resources managers to preserve an optimal exploit from these vital resources.

     

A Comparative Assessment of Metaheuristic Optimized Extreme Learning Machine and Deep Neural Network in Multi-Step-Ahead Long-term Rainfall Prediction for All-Indian Regions

Prediction of long-term rainfall patterns is a highly challenging task in the hydrological field due to random nature of rainfall events. The contribution of monthly rainfall is important in agriculture and hydrological tasks. This paper proposes two data-driven models, namely biogeography-based extreme learning machine (BBO-ELM) and deep neural network (DNN), to predict one, two, and three month-ahead rainfall over India (All-India and six other homogeneous regions). Three other data-driven models called ELM, genetic algorithm (GA)-based ELM, and particle swarm optimization (PSO)-based ELM are used to compare the performance of the proposed models. Firstly, partial autocorrelation function (PACF) is applied in all datasets to select the optimal number of lags for input to the models. Secondly, the wavelet-based data pre-processing technique is applied in selected optimal lags and feed to the proposed models for achieving higher prediction performance. To investigate the performance of proposed models, a non-parametric statistical test, Anderson–Darling’ Normality test, is performed in all India dataset. The wavelet-based proposed hybrid models show better prediction capability compared to optimal lag-based proposed models. This study shows the successful application of time-series data using proposed techniques (optimal lags-based BBO-ELM and wavelet-based DNN) in the hydrological field which may be used for risk mitigation from dreadful natural events.

     

Integrated Economic-Hydrologic Modelling for Groundwater Basin Management

This paper employs a modified penalty function and conjugate gradient based nonlinear programming algorithm to solve large integrated groundwater management problems.Discrete kernels are used to incorporate spatial and temporal distribution of the aquifer response due to external stresses. Significant improvements in computational speed and storage requirement are achieved owing to this unique conjugate gradient-discrete kernel combination. Simulations of the 'socially optimal' and 'open access' groundwater extraction schemes are presented as case studies of important policy implications.     

Cost-effective Approaches for Sustainable Groundwater Management in Alluvial Aquifer Systems

The consequences of unsustainable use of groundwater are becoming increasingly evident worldwide, particularly in developing nations. Groundwater management is a serious problem in many parts of the world. The prime concern is how to achieve groundwater sustainability. Artificial recharge techniques coupled with water harvesting hold a great promise for groundwater sustainability. The main intent of this paper is to highlight salient cost-effective and easy-to-use methods for augmenting groundwater resources in the alluvial hydrogeologic setting. Based on the intensive field investigation in a groundwater basin of Japan, three low-cost and easy-to-implement recharge techniques are suggested for alluvial aquifer systems, viz., augmentation of river flow, recharge through irrigation/drainage canals, and recharge from paddy fields. The source of recharge water for these recharge techniques is surface water supply. The efficacy of these methods or approaches has been demonstrated. It is emphasized that such cost-effective methods of groundwater recharge are key to sustainable groundwater management in both developing and developed countries. These methods must not be overlooked in the midst of sophisticated and highly expensive methods.     

Potential of Hybrid Data-Intelligence Algorithms for Multi-Station Modelling of Rainfall

Water Resources Management - One of the most challenging tasks in rainfall prediction is designing a reliable computational methodology owing the random and stochastic characteristics of...     

New Machine Learning Ensemble for Flood Susceptibility Estimation

Water Resources Management - Floods are among the most severe natural hazard phenomena that affect people around the world. Due to this fact, the identification of zones highly susceptible to...     

Estimating Pumping Rates and Identifying Potential Recharge Zones for Groundwater Management in Multi-Aquifers System

Agricultural, aquaculture, industrial and domestic activities have placed enormous demands for water, which sometimes results in the over-pumping and associated continually declining groundwater levels. This in turn has led to land subsidence and soil salination. Therefore, it’s important to understand the local pumping activities or the pumping rates in order to implement appropriate water management. The distribution of pumping rates varies spatially and temporally due to the availability of surface water and seasonality. In addition, to have correct estimate of the pumping rates, both the hydrology and geology should be consider. SWAT and MODFLOW are employed and run separately to acquire certain hydrologic components such as the recharge, boundary flow and change of aquifer storage in multi-aquifers. The water balance method (WBM) is then adopted to estimate pumping rates with these components. To validate the proposed model the results of WBM and the official records are compared. Besides, in view of the serious land subsidence occurred, artificial recharge is regarded as an effective tool to alleviate and mitigate the subsidence. Nevertheless, the location of conducting artificial recharge needs to be identified first. The potential recharge zones are assessed based on the simulated recharge rates from SWAT and the spatial distribution of hydrological characteristics of the unconfined aquifer. Ultimately, an optimal recharge zone will be suggested. The proposed methodology is proved capable of estimating the pumping rates and locating the potential recharge zone.     

Simulation-Optimization Modelling for Sustainable Groundwater Management in a Coastal Basin of Orissa, India

The Balasore coastal groundwater basin of Orissa in eastern India is under a serious threat of overdraft and seawater intrusion. Two optimization models were developed in this study for the efficient utilization of water resources in Balasore basin during non-monsoon periods: (a) a non-linear hydraulic management model for optimal pumpage, and (b) a linear optimization model for optimal cropping pattern in integration with a calibrated and validated groundwater flow simulation model. Based on the simulation-optimization modeling results, optimal pumping schedules, cropping patterns, and corresponding groundwater conditions are presented for three scenarios viz., wet, normal and dry years. It was found that optimal pumping schedules and corresponding cropping patterns differed significantly under the three scenarios, and the groundwater levels improved significantly under the optimal hydraulic conditions compared to the existing condition. In dry years, the groundwater levels under the present pumping pattern and the optimal pumpage indicated that the non-monsoon pumpage should not exceed the optimal pumpage in the absence of remedial measures in the basin. It is concluded that in order to ensure sustainable groundwater utilization in the basin, the optimal cropping pattern and pumping schedule should be adopted by the farmers.     

Evaluating Different Machine Learning Models for Runoff and Suspended Sediment Simulation

Water Resources Management - In the present study, prediction of runoff and sediment at Polavaram and Pathagudem sites of the Godavari basin was carried out using machine learning models such as...     

Machine Learning Utilization for Bed Load Transport in Gravel-Bed Rivers

Three data-driven techniques, namely artificial neural networks, adaptive-network-based fuzzy inference system, and symbolic regression based on genetic programming, are employed for the prediction of bed load transport rates in gravel-bed steep mountainous streams and rivers in Idaho (U.S.A.), and the potential of several input variables is investigated. The input combinations that were tested are based, mainly, on unit stream power, stream power, and shear stress, and exhibited similarly good performance, with respect to the machine learning technique used, accentuating the importance of the regression model. The derived models are robust, generalize very well in unseen data, and generated results superior to those of some of the widely used bed load formulae, without the need to set a threshold for the initiation of motion, and consequently avoid predicting erroneous zero transport rates.     

平顶山市地下水资源分布及开采潜力研究

研究区地下水分布广泛,资源量大,开发利用程度高,全区共有四种类型的地下水,即松散岩类孔隙水、碳酸盐岩类裂隙岩溶水、碎屑岩类裂隙孔隙水和基岩裂隙水,多年平均地下水资源总量82468.14万m3,可开采资源量72435.88万m3,地下水资源空间分布不均,含水层富水性差异明显,现状条件下地下水开采系数为0.85,处于采补基...     

Water Resources Modelling under Data Scarcity: Coupling MIKE BASIN and ASM Groundwater Model

The Water Framework Directive calls for strategic water resources planning at a catchment level, yet data and information are scarce in the areas where they are most needed: in the new EU Member States and Third Counties trying to assess the impact of EU environmental legislation in their water resources management policy. The research presented here proposes the coupling of a strategic scale water resources management simulation model (MIKE-Basin) and a finite difference groundwater model (ASM), as a tool to support decision making in data scarce environments. The models were applied in a particularly data scarce region, the Vrbas River basin, in Republic Srpska (RS) in Bosnia and Herzegovina (BiH) and the results are presented and discussed. It is argued that the approach adopted is valid and useful as an initial knowledge development and optioneering step, which can guide a national data collection exercise to support detailed modelling, and inform a strategic decision making process relevant to the application of the water framework directive.     

Anomaly Detection and Reliability Analysis of Groundwater by Crude Monte Carlo and Importance Sampling Approaches

In this paper, Crude Monte Carlo method and importance sampling are utilized to determine the reliability of long-term changes in groundwater level. Furthermore, different data analysis methods are used to determine the abnormal patterns and to investigate the cause of spatial variations of failure probability. For this purpose, three methods including robust covariance, one-class SVM, and Isolation Forest are applied to define the decision function. In the preliminary detection of the outliers, DFFITS and COOK measures are used to confirm the existence of abnormal plains in a two-dimensional space. The validity of prediction results is verified through the developed method of uncertain monitoring by selecting the most significant outlier points. In addition, the abnormal pattern detection methods are compared using non-random pattern discovery decision functions. The reliability analysis of groundwater is conducted during the two periods from 1994 to 2007 and 2008 to 2021. In the second period, parts of the eastern part of the northwest, central parts of the desert of Iran, and areas from west-southwest and east-south-east to other regions exposed to a lower probability of passing through the critical conditions. In contrast, the outcomes confirm the occurrence of drought with probability more than 80% for most of the plains. Eventually, the importance sampling method showed the closest relation in the correct distribution of the decision function. In contrast, due to the cluster shape and density of the outliers, the upper part of the decision function was determined with high certainty in the discovery of abnormal plains.     

Groundwater Modelling in Aquifers with highly Karstic and Heterogeneous Characteristics (KHC) in Palestine

Groundwater modelling is hindered by the lack of adequateinformation about the groundwater system and hence theneed for an interactive and efficient system for datapreparation and results analysis. Such a lack ofinformation usually necessitates the use of tediousiterative methodology within a sensitivity analysisscheme. The heterogeneous aquifer systems complicate theissue since more data is required to simulate the system.This study demonstrates the integrated approach to bridgethe gap between data handling and modelling. The karstcretaceous aquifer system (complex aquifer system) of theEastern Basin in the West Bank is used to illustrate thisapproach. The groundwater modelling approach integrates the outputsfrom different programs for data preparation andanalysis. These include (1) Groundwater Database (GWW)(2) Geographic Information System (GIS) (3) GroundwaterModelling System (GMS). In addition, the paper will summarize the data collectionefforts, problems faced and experience gained workingwith heterogeneous media. This involves linking theresults from various field investigations for groundwaterdevelopment programs in the West Bank.     

Groundwater Overexploitation Causing Land Subsidence: Hazard Risk Assessment Using Field Observation and Spatial Modelling

Hazard risk assessment of land subsidence is a complicated issue aiming at identifying areas with potentially high environmental hazard due to land subsidence. The methods of hazard risk assessment of land subsidence were reviewed and a new systematic approach was proposed in this study. Quantitative identification of land subsidence is important to the hazard risk assessment. Field observations using extensometers were used to determine assessment indexes and estimate weights of each index. Spatial modelling was also established in ArcGIS to better visualize the assessment data. These approaches then were applied to the Chengnan region, China as a case study. Three factors, thickness of the second confined aquifer, thickness of the soft clay and the annual recovery rate of groundwater level were incorporated into the hazard risk assessment index system. The weights of each index are 0.33, 0.17 and 0.5 respectively. The zonation map shows that the high, medium and low risk ranked areas for land subsidence account for 9.5?%, 44.7?% and 45.8?% of the total area respectively. The annual recovery rate of groundwater level is the major factor raising land subsidence hazard risk in approximately half of the study area.     

Potential Clogging and Dissolution Effects During Artificial Recharge of Groundwater Using Potable Water

To avoid the water quality deterioration that are caused by artificial recharge (AR) of groundwater, potable drinking water has been used as one of the source water for AR to control the side effects caused by the over-exploitation of groundwater. Chemical clogging problems can still be caused by certain chemical components, especially Fe and Al, and a lower concentration of these elements can cause a notable decrease in hydraulic conductivity at the top layer of the infiltration medium. Some components in AR source water can be obstructed by the clogging layer, leading to a change in water quality. The accumulation of total suspended solids (TSS) at the clogging layer can cause physical clogging and worsen the degree of chemical clogging. Although clogging and the related change in water quality were the dominant issues that affect the infiltration rate and health risks during the AR process, the dissolution of the aquifer matrix should also be taken into account. This dissolution contributed to not only the hydraulic conductivity of the infiltration medium but also the potential change in water quality during the aquifer recharge, storage and recovery processes.     

基于极限学习机的堤防工程多元风险指标评价方法

为提高堤防工程风险评价的准确性,提出了基于极限学习机的堤防工程多元风险指标评价方法。首先,综合考虑影响堤防风险的28个评价指标,利用层次分析法从预警系统、堤防工程系统、环境系统和社会经济系统这4个方面建立了堤防工程多元风险评价指标体系。接着,基于极限学习机算法对28个指标进行标准化处理及分级标准构建,以风险指标作为输入量,分级隶属度作为输出量,划分风险等级,量化评价指标,估计多元风险评价指标值和判断风险的严重程度。最后,依托鄱阳湖重点堤防——康山大堤,构建多元风险评价指标体系,运用极限学习机算法计算多元风险评价指标值。评价结果表明:康山大堤目前处于基本安全水平,符合康山大堤经过两次加固后的工程实际情况,并与其他方法进行对比,验证了提出方法的可靠性和有效性。该方法可拓展应用到其他重要水工结构工程风险评估中。